社会影响力分析——模型、方法和评价
Research
Cybersecurity—Review
Social In?uence Analysis:Models,Methods,and
Evaluation
Kan Li ?,Lin Zhang,Heyan Huang
School of Computer Science and Technology,Beijing Institute of Technology,Beijing 100081,China
a r t i c l e i n f o Article history:
Received 10December 2017Revised 5January 2018Accepted 8January 2018
Available online 16February 2018Keywords:
Social in?uence analysis Online social networks
Social in?uence analysis models In?uence evaluation
a b s t r a c t
Social in?uence analysis (SIA)is a vast research ?eld that has attracted research interest in many areas.In this paper,we present a survey of representative and state-of-the-art work in models,methods,and eval-uation aspects related to SIA.We divide SIA models into two types:microscopic and macroscopic models.Microscopic models consider human interactions and the structure of the in?uence process,whereas macroscopic models consider the same transmission probability and identical in?uential power for all users.We analyze social in?uence methods including in?uence maximization,in?uence minimization,?ow of in?uence,and individual in?uence.In social in?uence evaluation,in?uence evaluation metrics are introduced and social in?uence evaluation models are then analyzed.The objectives of this paper are to provide a comprehensive analysis,aid in understanding social behaviors,provide a theoretical basis for in?uencing public opinion,and unveil future research directions and potential applications.
ó 2018 THE AUTHORS. Published by Elsevier LTD on behalf of Chinese Academy of Engineering and Higher Education Press Limited Company. This is an open access article under the CC BY-NC-ND license
1.Introduction
Online social networks such as Weibo,Twitter,and Facebook provide valuable platforms for information diffusion among their users.During this process,social in?uence occurs when a person’s opinions,emotions,or behaviors are affected by other people [1].Thus,changes occur in an individual’s attitudes,thoughts,feelings,or behaviors as a result of interaction with other people or groups.Social in?uence analysis (SIA)is becoming an impor-tant research ?eld in social networks.SIA mainly studies how to model the in?uence diffusion process in networks,and how to propose an ef?cient method to identify a group of target nodes in a network [2].Studied questions include:Who in?uences whom;who is in?uenced;who are the most in?uential users,and so forth.SIA has important social signi?cance and has been applied in many ?elds.Viral marketing [3–10],online recommen-dation [11],healthcare communities [12–14],expert ?nding [15–17],rumor spreading [18],and other applications all depend on the social in?uence effect [19–21].Analyzing social in?uence can help us to understand peoples’social behaviors,provide the-oretical support for making public decisions and in?uencing pub-lic opinion,and promote exchanges and dissemination of various activities [22].
This paper provides a comprehensive view of SIA from the aspects of models,methods,and evaluation.To this end,we iden-tify the strengths and weaknesses of existing models and methods,as well as those of the evaluation of social in?uence.First,we review existing social in?uence models.Next,we summarize social in?uence methods.Finally,we analyze the evaluation of social in?uence.
The rest of this paper is organized as follows.In Section 2,we discuss SIA models.In Section 3,we analyze SIA methods,includ-ing in?uence maximization,in?uence minimization,?ow of in?u-ence,and individual in?uence.We then detail social in?uence evaluation in Section 4.Finally,we summarize the reviewed mod-els and methods of social in?uence,and discuss open questions.2.Social in?uence analysis models
SIA models have been widely studied in the literature.We clas-sify these models into two categories:microscopic and macro-scopic models.2.1.Microscopic models
Microscopic models focus on the role of human interactions,and examine the structure of the in?uence process.The two fre-quently used in?uence analysis models in this category are the independent cascade (IC)[23–25]and linear threshold (LT)
?Corresponding author.
E-mail address:likan@https://www.360docs.net/doc/664755022.html, (K.Li).
Engineering 4(2018)
40–46
Contents lists available at ScienceDirect
Engineering
[23,26]models.Since Kempe et al.[23]used these two models,they are still mainly used to assess social in?uence diffusion.2.1.1.The IC and LT models
The IC model.In a social network G ?eV ;E Twith a seed set S (S #V ),where V is the set of nodes and E is the set of edges,and S t (S t #V )is the set of nodes that are activated at step t (t P 0).At step t t1,every node v i 2S t can activate its out-neighbors v j 2V n [06i 6t S i with an independent probability P ij .The process ends when no node can be activated.Note that a node has only one chance to activate its out-neighbors after it has been activated,and the node remains an activated node after it is activated.
The LT model.In a social network G ?eV ;E T,the sum of the in?u-ence weights of all the neighbors of node v i meets P
v j 2N g
act i
w ij 61,
where w ij is the in?uence weight between node v i and its neighbor node v j ,and N g act i
is the set of its neighbor nodes activated by node
v i .Node v i randomly selects its own threshold h i ,uniformly chosen
from 0to 1.Only when the sum of the in?uence weights of its neigh-bor nodes exceeds this threshold will v i be activated.
2.1.2.Variations
For both the IC and LT models,it is usually necessary to run the Monte Carlo simulation in order to estimate a node’s in?uence for a suf?cient time period.However,this is time-consuming and unsuitable for large-scale social networks.Many researchers have proposed methods to improve the IC and LT models.Here,we divide these improvements into four categories:variations of the IC model,variations of the LT model,variations of both the IC and LT models,and models that differ from the IC or LT models and their variations.We only list representative works in this paper.
(1)Variations of the IC model.Some researchers have consid-ered time delay and time-critical constraints for in?uence diffu-sion.Chen et al.[27]extended the IC model and proposed an IC model with meeting events (IC-M model).In the IC-M model,the activated node has the probability to meet the inactive https://www.360docs.net/doc/664755022.html,-pared with the general IC model,the calculation results from this model are closer to the actual situation,although the calculation of time consumption is slightly above that of the IC model.Feng et al.[28]incorporated novelty decay into the IC model.Based on previous studies,they found that repeated exposures have diminishing in?uence on users.Therefore,they developed a prop-agation path-based algorithm to assess the in?uence spread of seed nodes.There are two values on each edge of a social network:in?uence probability and expected in?uence delay time.Mohamadi-Baghmolaei et al.[29]considered important time and trust factors,and proposed the trust-based latency-aware indepen-dent cascade (TLIC)model.This is the ?rst time trust has been studied in a classic IC model.In the TLIC model,a node can change its state (i.e.,as active or inactive)with different probabilities for a trusted neighbor node than for a distrusted neighbor.Budak et al.[30]introduced the multi-campaign IC model,which models the diffusion of two cascades evolving simultaneously.They studied the notion of competing campaigns,in which the good campaign counteracts the effect of a bad campaign in a social network.
(2)Variations of the LT model.Liu et al.[31]considered the containment of competitive in?uence diffusion in social networks,and extended the LT model to construct the diffusion-containment (D-C)model.The traditional LT model is not applicable to that a situation concerns both the diffusion and the containment of the in?uence.In the D-C model,the state of a node is described by the activation probability;each node is only in?uenced by a neigh-bor with a higher probability,and the sum of the probabilities of possible node states is not greater than 1.Borodin et al.[32]
analyzed the competitive in?uence diffusion of different models based on the LT model.
(3)Variations of both the IC and LT models.Mohammadi et al.[33]considered the time delay and proposed two diffusion models:the delayed independent cascade (DIC)model and the delayed lin-ear threshold (DLT)model.In these two models,nodes have three states:active,inactive,and latent active.To pass from an inactive state to an active state,a node must pass through the middle pro-cess of being in a latent active https://www.360docs.net/doc/664755022.html,pared with the traditional IC and LT models,the effect of in?uence diffusion in this model is better.However,because more than one state is possible,the cal-culation complexity is relatively high.In the IC and LT models,information diffusion is treated as a series of node state changes that occur in a synchronous way.However,the actual diffusion takes place in an asynchronous way,and the time stamps of the observed data are not evenly spaced out.Some models relax the synchronicity assumption of traditional IC and LT models and extend these two models to make the state change asyn-chronously.Saito et al.[34]proposed the asynchronous exten-sions:asynchronous IC (AsIC)and asynchronous LT (AsLT)models.Their learning algorithm can estimate the in?uence degrees of nodes in a network from relatively few information dif-fusion results,which avoids the over?tting problem.Guille and Hacid [35]also presented an asynchronous model—time-based AsIC model—to model the diffusion process.Other studies [30,32,36–44]have improved the classical models in speci?c direc-tions.Fan et al.[38]introduced two models:the opportunistic one-activate-one (OPOAO)model,in which each person can only com-municate with another person simultaneously;and the determin-istic one-activate-many (DOAM)model,which has a mechanism that is similar to an information broadcast procedure.For the OPOAO model,they used the classical greedy algorithm to produce a (1à1=e)approximation ratio;for the DOAM model,they pro-posed a set cover-based greedy (SCBG)algorithm to achieve an O eln n T(n is the number of vertices)factor solution.The transmis-sion probabilities in these two models are the same for all users.Galam [45]proposed a model that investigates opinion dynamics,followed by Lee et al.[39],who proposed a new scheme to improve Galam’s model.Nevertheless,these models are con?ned to word-of-mouth information-exchanging process.
(4)Models that differ from the IC or LT models and their variations.Some models are different from the IC or LT models and their variations,and have solved information in?uence diffu-sion from a new point of view.Lin et al.[46]proposed a data-driven model to maximize the expected in?uence in the long run using meta-learning concepts.However,this model needs large amounts of data,and the accuracy of its results requires fur-ther improvement.Golnari et al.[47]proposed a heat conduction (HC)model.It considers a non-progressive propagation process,and is completely different from the previous IC or LT models,which only consider the progressive propagation process.In the HC model,the in?uence cascade is initiated from a set of seeds and arbitrary values for other nodes.Wang et al.[48]studied emotion in?uence in large-scale image social networks,and pro-posed an emotion in?uence model.They designed a factor graph model to infer emotion in?uence from images in social networks.Gao [49]proposed a read-write (RW)model to describe the detailed processes of opinion forming,in?uence,and diffusion.However,there are three main issues that this model needs to consider further:the many parameters of the model that must be inferred,the proper collection of datasets about opinion in?u-ence and diffusion,and the evaluation metrics that are suitable for this task.
Whether a classical IC or LT model or a new model,the ultimate goal for a model of social in?uence is to achieve faster computing speed and more accurate results,while using less memory.
K.Li et al./Engineering 4(2018)40–4641
However,these microscopic models are lacking in some major details.Firstly,because of the difference in characteristics (i.e.,educational background and individual consciousness),differ-ent people identify with the same information to different degrees. Secondly,different individuals have different capabilities to in?u-ence other users;also,spreaders with different degrees of author-ity have different impacts on their neighbors and on society. Thirdly,the probability that an individual will transmit a piece of information to others is not constant,and should depend on the individual’s attraction to the information.Moreover,these models use a binary variable to record whether an individual becomes infected.In addition,they assume that once an individual is infected,the individual never changes its state;however,this assumption does not re?ect the realistic smoothness of the transi-tion of individuals from one state to another.
2.2.Macroscopic models
Macroscopic models consider all users to have the same attrac-tion to information,the same transmission probability,and identi-cal in?uential power.However,since macroscopic models do not take individuals into account,the accuracy of these models’results is lower.To improve such models,therefore,the differences between individuals should be considered.Macroscopic models divide nodes into different classes(i.e.,states)and focus on the state evolution of the nodes in each class.The percentage of nodes in each class is expressed by simple differential equations.Epi-demic models are the most common models that are used to study social in?uence from a macroscopic perspective.These models were mainly developed to model epidemiological processes.How-ever,they neglect the topological characteristics of social net-works.The percentage of nodes in each class is computed by mean-?eld rate equations,which are too simple to depict such a complex evolution accurately.Daley and Kendall[50]analyzed the similarity between the diffusion of an infectious disease and the dissemination of a piece of information,and proposed the clas-sic Daley–Kendall model.Since then,researchers have improved these epidemic models in general to overcome their weaknesses. Refs.[50–53]consider the topological characteristics of underlying networks in their methods.However,these scholars have ignored the impact of human behaviors in the in?uence diffusion process.
Researchers have recently begun to consider the role of human behaviors and different mechanisms during information in?uence diffusion[54–60].Zhao et al.[59,60]proposed the susceptible–in fected–hibernator–removed model as an extension from the classi-cal susceptible–infected–removed(SIR)model in order to incorpo-rate the forgetting and remembering mechanism;they investigated this problem in homogeneous and inhomogeneous networks.Wang et al.[54]proposed a variation of the rumor-diffusion model in online social networks that considers negative or positive social reinforcements in the acceptant probability model.They analyzed how social reinforcement affects the spread-ing rate.Wang et al.[55]proposed an SIR model by introducing the trust mechanism between nodes.This mechanism can reduce the ultimate rumor size and the velocity of rumor diffusion.Xia et al.
[56]presented a modi?ed susceptible–exposed–infected–removed (SEIR)model to discuss the impact of the hesitating mechanism on the rumor-spreading model.They took into account the attractive-ness and fuzziness of the contents of rumors,and concluded that the more clarity a rumor has,the smaller its effects will be.Su et al.[57]proposed the microblog-susceptible–infected–removed model for information diffusion by explicitly considering users’incomplete reading behavior.In Ref.[58],motivated by the work in Refs.[56,57],Liu et al.extended the model in Ref.[56]and pro-posed a new SEIR model on heterogeneous networks in order to study the diffusion dynamics in a microblog.3.Social in?uence analysis methods
SIA methods are used to solve the sub-problems of social net-work in?uence analysis,such as in?uence maximization,in?uence minimization,?ow of in?uence,and individual in?uence.All these problems involve in?uence diffusion,so the same in?uence model can apply to all of them in some cases.However,the ultimate goal of using the model is different for each problem.
3.1.In?uence maximization
In?uence maximization requires?nding the most in?uential group of members in social networks.Kempe et al.[23]formulated this problem.Given a directed graph with users as nodes,edge weights that re?ect the in?uence between users,and a budget/ threshold number k,the purpose of in?uence maximization is to ?nd k nodes in a social network,so that the expected spread of the in?uence can be maximized by activating these nodes.This is a discrete optimization problem that is non-deterministic polynomial-time(NP)-hard for both the IC and LT models.In?u-ence maximization is the most widely studied problem in SIA.Les-kovec et al.[10]and Rogers[20]studied this problem as an algorithmic problem,and proposed some probabilistic methods. In?uence maximization must achieve fast calculation,high accu-racy,and low storage capacity.Most of the algorithms that con-sider the differences between individuals are based on greedy algorithms or heuristic algorithms.
3.1.1.Greedy algorithms
Greedy algorithms‘‘greedily”select the active node with the maximum marginal gain toward the existing seeds in each itera-tion.The study of greedy algorithms is based on the hill-climbing greedy algorithm,in which each choice can provide the greatest impact value of the node using the local optimal solution to approximate the global optimal solution.The advantage of this algorithm is that the accuracy is relatively high,reaching (1à1=eàe)for any e>0.However,the algorithm has high com-plexity and high execution time,so the ef?ciency is relatively poor. Kempe et al.[23]were the?rst to establish the in?uence of the maximum re?nement as a discrete optimization problem,and pro-posed a greedy climbing approximation algorithm.Leskovec et al.
[61]proposed a greedy optimization method,the cost-effective lazy forward(CELF)method.Chen et al.[27]put forward new greedy algorithms,the NewGreedy and MixGreedy methods.Zhou et al.[62]proposed the upper bound-based lazy forward(UBLF) algorithm to discover top-k in?uential nodes.They established new upper bounds in order to signi?cantly reduce the number of Monte Carlo simulations in greedy algorithms,especially at the ini-tial step.Some of the algorithms that are used to study the differ-ences between individuals are based on these greedy algorithms.
3.1.2.Heuristic algorithms
Due to the high computational complexity of greedy algorithms, many excellent heuristic algorithms have been proposed to reduce the solution-solving time and pursue higher algorithm ef?ciency. These heuristic algorithms iteratively select nodes based on a speci?c heuristic,such as degree or PageRank,rather than comput-ing the marginal gain of the nodes in each iteration.Their draw-back is that the accuracy is relatively low.The most basic heuristic algorithms are the Random,Degree,and Centrality heuristic algorithms proposed by Kempe et al.[23].Based on the degree of the heuristic algorithm,Chen et al.[41]proposed a heuristic algorithm for the IC model:DegreeDiscount.They then proposed a new heuristic algorithm,pre?x excluding maximum in?uence arborescence(PMIA)[63].For the LT model,Chen et al.
42K.Li et al./Engineering4(2018)40–46
[63]proposed local directed acyclic graph(LDAG)heuristic algo-rithm.Of course,other heuristic algorithms are also based on these heuristic algorithms,such as SIMPATH[64]and IRIE[65].Borgs et al.[66]made a theoretical breakthrough and presented a quasi-linear time algorithm for in?uence maximization under the IC model.Tang et al.[67]proposed a two-phase in?uence maximiza-tion(TIM)algorithm for in?uence maximization.The expected time of the algorithm is O?ekt‘TentmTlog n=e2 ,and it returns a (1à1=eàe)approximate solution with at leaste1ànà‘Tprobability.
The time complexity of the abovementioned algorithms is shown in Table1[23,41,61,63–67].
Here,we describe some representative studies that take into account individual differences or the user’s own attributes.Li et al.[68]considered the behavioral relationships between humans(i.e.,regular and rare behaviors)in order to simulate the effects of heterogeneous social networks to solve the in?uence maximization problem.They proposed two entropy-based heuris-tic algorithms to identify the communicators in the network,and then maximized the impact of the propagation.Although the entropy-based heuristic performance is better than those of Degree [23]and DegreeDiscount[41],this method is still based on heuris-tic ideas,so its accuracy is inadequate.Subbian et al.[69]proposed the individual social value:Existing in?uence maximization algo-rithms cannot model individual social value,which is usually the real motivation for nodes to connect to each other.Subbian et al.
[69]used the concept of social capital to propose a framework in which the social value of the network is calculated by the number of bindings and bridging connections.The performance of the algo-rithm is better than that of PMIA[63],PageRank,and the weighted degree method.Li et al.[70]proposed a conformity-aware cascade (C2)model and a conformity-aware greedy algorithm to solve the in?uence maximization problem.This algorithm works well when applied to the distributed platform.However,because it is based on the greedy algorithm and considers the conformity-aware cal-culation,the complexity of the calculation still requires improve-ment.Lee and Chung[71]formulated the in?uence maximization problem as a query processing problem for distinguishing speci?c users from others.Since in?uence maximization query processing is NP-hard,and since solving the objective function is also NP-hard,they focused on how to approximate optimal seeds ef?-ciently.Node features are always overlooked when estimating the impact of different users.Deng et al.[72]addressed in?uence maximization while considering this factor.They presented three quantitative measures to respectively evaluate node features:user activity,user sensitivity,and user af?nity.They combined node features into users’static effects,and then used the continuous exponential decay function to convert the strength of a user’s dynamic in?uence between two adjacent users.They also pro-posed the credit distribution with node features(CD-NF)model, which rede?nes the credit,and designed the greedy algorithm with node features(GNF)based on the CD-NF model.
When considering a certain class or in?uence maximization,the individual’s attributes will have a great impact on the results; therefore,individual differences or the user’s attributes need to be taken into account.From the descriptions above of the studies on individual differences,it is clear that most studies are based on the greedy algorithm in order to improve the accuracy.How-ever,heuristic algorithms will be improved by a pursuit of high ef?ciency and low complexity,because the greedy algorithm is computationally complex.In order to reduce the running time,a considerable amount of work will be carried out on a distributed platform.Due to the individual differences,the computational complexity will be further increased compared with the original. So,for the time being,it is necessary to improve the complexity of the calculation and the accuracy of the results.
3.2.In?uence minimization
Assuming that the negative information propagates in a net-work G?eV;ETwith the initially infected node set S#V,the goal of in?uence minimization is to minimize the number of?nal infected nodes by blocking k nodes(or vertices)of the set D#V, where k((|V|)is a given constant.It can be expressed as the fol-lowing optimization problem:
D??arg min
D#V;j D j6k
reS V n D
jTe1Twhere reS V n D
jTdenotes the in?uence of set S when nodes in the set D are blocked.
From this notion[73],we know that in?uence minimization is a dual problem of in?uence maximization.In?uence minimization mainly used to curb rumors,monitor public opinion,and so on. Yao et al.[73]proposed a method to minimize adverse effects in the network by preventing a limited number of nodes from the perspective of the topic model.When rumors and other adverse events occur in social networks and some users are already infected,the purpose of this model is to minimize the number of ?nal infected users.Wang et al.[74]proposed a dynamic rumor in?uence minimization model with user experience.This model minimizes the in?uence of rumors(i.e.,the number of users who accept and send rumors)by preventing part of the nodes from acti-vating.Groeber et al.[75]designed a general framework of social in?uence inspired by the social psychological concept of cognitive dissonance,in which individuals minimize the preconditions for disharmony arising from disagreements with neighbors in a given social network.Chang et al.[76]explored the?rst solution to esti-mate the probability of successfully bounding the infected count below the out-of-control threshold;this can be logically mapped to outbreak risk,and can allow authorities to adaptively adjust the intervention cost to meet necessary risk control.They then pro-posed an in?uence minimization model to effectively prevent the proliferation of epidemic-prone diseases on the network.
3.3.Flow of in?uence
When information spreads,it is accompanied by in?uence?ow [77].In recent years,the?ow of in?uence methods has been stud-ied by many researchers.Subbian et al.[78]proposed a?ow
Table1
The time complexity of different algorithms.
Algorithm Time complexity
Hill-climbing greedy[23]OeknRmT
CELF[61]OeknRmT
NewGreedyIC[41]OekRmT
NewGreedyWC[41]OekRTmT
MixGreedyIC[41]OekRmT
MixGreedyWC[41]OekRTmT
DegreeDiscountIC[41]Oek log ntmT
PMIA[63]O?nt i htkn o h n i hen i htlog nT
LDAG[63]Oe‘vtn log‘vT
SIMPATH[64]OekmnT
IRIE[65]OekmnT
Quasilinear time algorithm[66]O?k eà2emtnTlog n
TIM[67]O?ekt‘TemtnTlog n=e2
n:number of vertices in G;m:number of edges in G;k:number of seeds to be
selected;R:number of rounds of simulations;T:number of iterations;t i h,n o h,n i h:
constants decided by h;h:the in?uence threshold;n i h?max v2V fj MIIAev;hTjg;
n o h?max v2V fj MIOAev;hTjg;MIIAev;hT=MIOAev;hT:the maximum in?uence in
arborescence/out arborescence of a node v;t i h:the maximum running time to
compute MIIAev;hT;‘:number of communities in the network;e:any constant
larger than0;‘v:the volume of LDAGev;hT.
K.Li et al./Engineering4(2018)40–4643
pattern mining approach with the condition of speci?c?ow valid-ity constraints.Kutzkov et al.[79]proposed a streaming method called STRIP to compute the in?uence strength along each link of a social network.Teng et al.[80]examined real-world information ?ows in various platforms,including the American Physical Soci-ety,Facebook,Twitter,and LiveJournal,and then leveraged the behavioral patterns of users to construct virtual information in?u-ence diffusion processes.Chintakunta and Gentimis[81]discussed the relationships between the topological structures of social net-works and the information?ows within them.However,unlike microblogging platforms,most social networks cannot provide suf-?cient context to mine the?ow pattern.
3.4.Individual in?uence
Individual in?uence is a relatively microscopic assessment that models the in?uence of a user on other users or on the whole social network.Chintakunta and Gentimis[81]proposed a method called SoCap to?nd in?uencers in a social network.They modeled in?u-encer?nding in a social network as a value-allocation problem,in which the allocated value represents the individual social capital. Subbian et al.[82]proposed an approach to identify in?uential agents in open multi-agent systems using the matrix factorization method to measure the in?uence of nodes in a network.Liu et al.
[83]presented the trust-oriented social in?uence(TOSI)method, which considers social contexts(i.e.,social relationships and social trust between participants)and preferences in order to assess indi-vidual in?uence.The TOSI method greatly outperforms SoCap in terms of effectiveness,ef?ciency,and robustness.Deng et al.[84] incorporated the time-critical aspect and the characteristics of the nodes when evaluating the in?uence of different users.The results showed that their approach is ef?cient and reasonable for identifying seed nodes,and its prediction of in?uence spread is more accurate than that of the original method,which disregards node features in the diffusion process.
In conclusion,considering more comprehensive user character-istics and user interaction information results in higher result accuracy.
4.Social in?uence evaluation
4.1.In?uence evaluation metrics
Running time is a very intuitive measure of model ef?ciency that is easy to calculate.In general,under the same conditions,the faster a model runs,the better it is.However,the traditional greedy algo-rithm calculates the range of in?uence spreading for a given node set by a large number of repeated Monte Carlo simulations,result-ing in a considerable running time.Especially in the face of current large-scale social networks,the existing algorithms cannot meet the application requirements for ef?ciency.Therefore,running time is an important measure of social in?uence evaluation.
Since the in?uence-spreading problem is NP-hard,it is dif?cult to obtain an optimal solution of the objective function.Most of the existing algorithms rely on monotonicity and submodularity of the function to achieve(1à1=e)approximation[23].However, attempts to achieve a higher approximation ratio have never stopped.Zhu et al.[85]proposed semide?nite-based algorithms in their model considering in?uence transitivity and limiting prop-agation distance.
Another metric is the number of Monte Carlo calls.Because there is no way to obtain an optimal solution,a Monte Carlo sim-ulation is usually used to estimate the real value.Existing greedy-based algorithms demand heavy Monte Carlo simulations of the spread functions for each node at the initial step,greatly reducing the ef?ciency of the models.The UBLF algorithm proposed by Zhou et al.[62]can reduce the number of Monte Carlo simulations of CELF method by more than95%,and achieved a speedup of2–10 times when the seed set is small.
The expected spread indicates the number of nodes that the seed set can ultimately affect—and the larger the better.There are many applications in real-life scenarios where the in?uence spread needs to be maximized as much as possible.Typical exam-ples of such applications are marketing and advertising.In both applications,the?nal expected spread represents the bene?ts of product promotion or the pro?tability of product.Therefore, exploring high expected spread algorithms is an important prob-lem for SIA.
Robustness refers to the characteristics of a certain parameter perturbation(i.e.,structure and size)that is used to maintain some other performances.Both Jung et al.[65]and Liu et al.[83]men-tioned robustness in their algorithms.The IRIE algorithm proposed by Jung et al.[65]is more robust and stable in terms of running time and memory usage across various density networks and cas-cades of different sizes.The experimental results showed that the IRIE algorithm runs two orders of magnitude faster than existing methods such as PMIA[63]on a large-scale network,and only uses part of the memory.The TOSI evaluation method proposed by Liu et al.[83]shows superior performance in terms of robustness over the state-of-the-art SoCap[81].
Scalability refers to the ability to continuously expand or enhance the functionality of the system with minimal impact on existing systems.In social networks,scalability usually refers to the ability to expand from a small-scale network to a large-scale network.It is a common indicator used to evaluate the quality of a model.Due to the complexity of the algorithm and the long run-ning time,the current solution algorithms only apply to small and medium-sized social networks with nodes below one million. Given today’s large-scale social networks,in?uence analysis algo-rithms with good scalability must be designed to deal with the challenges posed by massive social network data.
4.2.Social in?uence evaluation model
The evaluation of social in?uence is a complex process.As a subjective attribute,a social relationship has many characteristics, including dynamics,event disparity,asymmetry,transitivity,etc. In social networks,frequent user interaction and changes in the structure of the network make the evaluation of social in?uence more dif?cult.The literature contains a few studies on the social in?uence evaluation model.He et al.[86]designed an in?uence-measuring model on the theme of online complaints;based on the entropy weight model,this model monitors and analyzes the static and dynamic properties of complaint information in real time.Enterprises can use this model to manage online group com-plaints.Wang et al.[12]proposed a?ne-grained feature-based social in?uence(FBI)evaluation model,which explores the impor-tance of a user and the possibility of a user impacting others.They then designed a PageRank algorithm-based social in?uence adjust-ment model by identifying the in?uence contributions of friends. The FBI evaluation model can identify the social in?uences of all users with much less duplication(less than7%with the model), while having a larger in?uence spread with top-k in?uential users; it was evaluated on three datasets:HEPTH[87],DBLP[88],and ArnetMiner[89].
5.Conclusions and future work
In this paper,we survey state-of-the-art research on SIA from the aspects of in?uence models,methods,and evaluation.We also
44K.Li et al./Engineering4(2018)40–46
analyze the strengths and weaknesses of current models and methods.Throughout our study,we unveil future research direc-tions and potential applications.
In social in?uence models,we distinguish two types of models: microscopic and macroscopic models.Microscopic models con-sider human interactions and the structure of the in?uence pro-cess.Macroscopic models consider the same transmission probability and identical in?uential power for all users.In future, macroscopic models should focus on how to consider human behaviors and different mechanisms during information diffusion. Although many researchers have put considerable effort to improving the classical models and proposing new models from different perspectives—such as by adding constraints into models and incorporating competitive in?uence diffusion—there is still room for improvement.
In most existing models,a person is in?uenced only by the other person in a monoplex network,and the in?uence diffusion processes are independent.However,in real life,people often com-municate with others in multiplex networks.In this situation, social in?uence occurs in multiplex networks,and information in?uence among different monoplex networks encounters cooper-ation and competition.The question of how to model information in?uence in multiplex networks is a valuable research topic.In addition,the question of how to compute information in?uence over time in dynamic networks should be studied.In most exper-iments,datasets cover up to about100000nodes,so the issues inherent in applying social network analysis-related issues to mas-sive datasets(which may include millions or tens of millions of nodes,or even more)require study.In short,there is still room for research in extending SIA models to address perceived limita-tions such as ef?ciency and scalability.
Acknowledgements
This research was supported in part by the National Basic Research Program of China(2013CB329605).The authors thank Lingling Li and Junying Shang for their helpful discussions and comments.
Compliance with ethics guidelines
Kan Li,Lin Zhang,and Heyan Huang declare that they have no con?ict of interest or?nancial con?icts to disclose.
References
[1]Travers J,Milgram S.The small world problem.Psychol Today1967;1:61–7.
[2]Chen W,Lakshmanan LV,Castillo https://www.360docs.net/doc/664755022.html,rmation and in?uence propagation in
social networks.San Rafael:Morgan&Claypool;2013.
[3]Freeman LC.A set of measures of centrality based on betweenness.Sociometry
1977;40(1):35–41.
[4]Baas F.A new product growth model for consumer durables.Manage Sci
1969;15(5):215–27.
[5]Brown JJ,Reingen PH.Social ties and word-of-mouth referral behavior.J
Consum Res1987;14(3):350–62.
[6]Mahajan V,Muller E,Bass FM.New product diffusion models in marketing:A
review and directions for research.J Mark1990;54(1):1–26.
[7]Domingos P,Richardson M.Mining the network value of customers.In:
Proceedings of the7th ACM SIGKDD International Conference on Knowledge Discovery and Data Mining;2001Aug26–29;San Francisco,CA,USA;2001.p.
57–66.
[8]Goldenberg J,Libai B,Muller E.Talk of the network:a complex systems look at
the underlying process of word-of-mouth.Mark Lett2001;12(3):211–23. [9]Richardson M,Domingos P.Mining knowledge-sharing sites for viral
marketing.In:Proceedings of the8th ACM SIGKDD International Conference on Knowledge Discovery and Data Mining;2002Jul23–26;Edmonton,AB, Canada;2002.p.61–70.
[10]Leskovec J,Adamic LA,Huberman BA.The dynamics of viral marketing.J ACM
Trans Web2007;1(1):5.[11]Pálovics R,Benczúr AA,Kocsis L,Kiss T,FrigóE.Exploiting temporal in?uence
in online recommendation.In:Proceedings of the8th ACM Conference on Recommender Systems;2014Oct6–10;Foster City,CA,USA;2014.p.273–80.
[12]Wang G,Jiang W,Wu J,Xiong Z.Fine-grained feature-based social in?uence
evaluation in online social networks.IEEE Trans Parallel Distrib Syst2014;25
(9):2286–96.
[13]Christakis NA,Fowler JH.The spread of obesity in a large social network over
32years.N Engl J Med2007;357(4):370–9.
[14]Fowler JH,Christakis NA.Dynamic spread of happiness in a large social
network:longitudinal analysis over20years in the Framingham Heart Study.
Br Med J2009;338(7685):23–7.
[15]Franks H,Grif?ths N,Anand SS.Learning in?uence in complex social networks.
In:Proceedings of the2013International Conference on Autonomous Agents and Multi-agent Systems;2013May6–10;Saint Paul,MN,USA;2013.
p.447–54.
[16]Dong W,Pentland A.Modeling in?uence between experts.In:Proceedings of
the ICMI2006and IJCAI2007International Conference on Arti?cial Intelligence for Human Computing;2006Nov3;Banff,AB,Canada;2007.p.
170–89.
[17]Tang J,Sun J,Wang C,Yang Z.Social in?uence analysis in large-scale networks.
In:Proceedings of the15th ACM SIGKDD International Conference on Knowledge Discovery and Data Mining;2009Jun28–Jul1;Paris,France;
2009.p.807–16.
[18]He Z,Cai Z,Wang X.Modeling propagation dynamics and developing
optimized countermeasures for rumor spreading in online social networks.
In:Proceedings of the2005IEEE35th International Conference on Distributed Computing Systems;2015Jun29–Jul2;Columbus,OH,USA;2015.p.205–14.
[19]Katz E,Lazarsfeld PF.Personal in?uence:The part played by people in the?ow
of mass communications.New York:Free Press;1965.
[20]Rogers EM.Diffusion of innovations.5th ed.New York:Free Press;2003.
[21]Keller E,Berry J.The in?uentials:one American in ten tells the other nine how
to vote,where to eat,and what to buy.New York:Free Press;2003.
[22]Peng S,Yang A,Cao L,Yu S,Xie D.Social in?uence modeling using information
theory in mobile social networks.Inf Sci2017;379:146–59.
[23]Kempe D,Kleinberg J,Tardosé.Maximizing the spread of in?uence through a
social network.In:Proceedings of the9th ACM SIGKDD International Conference on Knowledge Discovery and Data Mining;2003Aug24–27;
Washington,DC,USA;2003.p.137–46.
[24]Leskovec J,Mcglohon M,Faloutsos C,Glance NS,Hurst M.Patterns of cascading
behavior in large blog graphs.In:Proceedings of the2007SIAM International Conference on Data Mining;2007Apr26–28;Minneapolis,MN,USA;2007.
[25]Gruhl D,Guha R,Liben-Nowell D,Tomkins https://www.360docs.net/doc/664755022.html,rmation diffusion through
blogspace.In:Proceedings of the13th International Conference on World Wide Web;2004May17–20;New York,NY,USA;2004.p.491–501.
[26]Granovetter M.Threshold models of collective behavior.Am J Sociol1978;83
(6):1420–43.
[27]Chen W,Lu W,Zhang N.Time-critical in?uence maximization in social
networks with time-delayed diffusion process.In:Proceedings of the26th AAAI Conference on Arti?cial Intelligence;2012Jul22–26;Toronto,ON, Canada;2012.p.592–8.
[28]Feng S,Chen X,Cong G,Zeng Y,Chee YM,Xiang Y.In?uence maximization with
novelty decay in social networks.In:Proceedings of the28th AAAI Conference on Arti?cial Intelligence;2014Jul27–31;Québec City,QC,Canada;2014.p.
37–43.
[29]Mohamadi-Baghmolaei R,Mozafari N,Hamzeh A.Trust based latency aware
in?uence maximization in social networks.J Eng App Artif Intell2015;41
(C):195–206.
[30]Budak C,Agrawal D,Abbadi AE.Limiting the spread of misinformation in social
networks.In:Proceedings of the20th International Conference on World Wide Web;2011Mar28–Apr1;Hyderabad,India;2011.p.665–74.
[31]Liu W,Yue K,Wu H,Li J,Liu D,Tang D.Containment of competitive in?uence
spread in social networks.Knowl Base Syst2016;109(C):266–75.
[32]Borodin A,Filmus Y,Oren J.Threshold models for competitive in?uence in
social networks.In:Proceedings of the6th International Conference on Internet and Network Economics;2010Dec13–17;Stanford,CA,USA;2010.p.
539–50.
[33]Mohammadi A,Saraee M,Mirzaei A.Time-sensitive in?uence maximization in
social networks.J Inf Sci2015;41(6):765–78.
[34]Saito K,Ohara K,Yamagishi Y,Kimura M,Motoda H.Learning diffusion
probability based on node attributes in social networks.In:Proceedings of the 19th International Conference on Foundations of Intelligent Systems;2011Jun 28–30;Warsaw,Poland;2011.p.153–62.
[35]Guille A,Hacid H.A predictive model for the temporal dynamics of
information diffusion in online social networks.In:Proceedings of the21st International Conference on World Wide Web;2012Apr16–20;Lyon,France;
2012.p.1145–52.
[36]Bharathi S,Kempe D,Salek https://www.360docs.net/doc/664755022.html,petitive in?uence maximization in social
networks.In:Proceedings of the3rd International Conference on Internet and Network Economics;2007Dec12–14;San Diego,CA,USA;2007.p.
306–11.
[37]Carnes T,Nagarajan C,Wild SM,Zuylen AV.Maximizing in?uence in a
competitive social network:A follower’s perspective.In:Proceedings of the 9th International Conference on Electronic Commerce;2007Aug19–22;
Minneapolis,MN,USA;2007.p.351–60.
[38]Fan L,Lu Z,Wu W,Thuraisingham B,Ma H,Bi Y.Least cost rumor blocking in
social networks.In:Proceedings of the2013IEEE33rd International
K.Li et al./Engineering4(2018)40–4645
Conference on Distributed Computing Systems;2013Jul8–11;Philadelphia, PA,USA;2013.p.540–9.
[39]Lee W,Kim J,Yu H.CT-IC:Continuously activated and time-restricted
independent cascade model for viral marketing.In:Proceedings of the2012 IEEE12th International Conference on Data Mining;2012Dec10–13;Brussels, Belgium;2013.p.960–5.
[40]Kostka J,Oswald YA,Wattenhofer R.Word of mouth:Rumor dissemination in
social networks.In:Proceedings of the15th International Colloquium on Structural Information and Communication Complexity;2008Jun17–20;
Villars-sur-Ollon,Switzerland;2008.p.185–96.
[41]Chen W,Wang Y,Yang S.Ef?cient in?uence maximization in social networks.In:
Proceedings of the15th ACM SIGKDD International Conference on Knowledge Discovery and Data Mining;2009Jun28–Jul1;Paris,France;2009.p.199–208.
[42]Wang Y,Wang H,Li J,Gao H.Ef?cient in?uence maximization in weighted
independent cascade model.In:Proceedings of the21st International Conference on Database Systems for Advanced Applications;2016Mar27–30;Dallas,TX,USA;2016.p.49–64.
[43]Pathak N,Banerjee A,Srivastava J.A generalized linear threshold model for
multiple cascades.In:Proceedings of the2010IEEE International Conference on Data Mining;2010Dec13–17;Sydney,Australia;2010.p.965–70. [44]Bharathi S,Kempe D,Salek https://www.360docs.net/doc/664755022.html,petitive in?uence maximization in social
networks.In:Proceedings of the3rd International Workshop on Web and Internet Economics;2007Dec12–14;San Diego,CA,USA;2007.p.306–11.
[45]Galam S.Modelling rumors:the no plane Pentagon French hoax case.Phys A
2003;320:571–80.
[46]Lin SC,Lin SD,Chen MS.A learning-based framework to handle multi-round
multi-party in?uence maximization on social networks.In:Proceedings of the 21st ACM SIGKDD International Conference on Knowledge Discovery and Data Mining;2015Aug10–13;Sydney,Australia;2015.p.695–704.
[47]Golnari G,Asiaee A,Banerjee A,Zhang ZL.Revisiting non-progressive in?uence
models:Scalable in?uence maximization.In:Proceedings of the31st Conference on Uncertainty in Arti?cial Intelligence;2015Jul12–16;
Amsterdam,The Netherlands;2015.
[48]Wang X,Jia J,Tang J,Wu B,Cai L,Xie L.Modeling emotion in?uence in image
social networks.IEEE Trans Affect Comp2015;6(3):286–97.
[49]Gao D.Opinion in?uence and diffusion in social network.In:Proceedings of
the35th International ACM SIGIR Conference on Research and Development in Information;2012Aug12–16;Portland,OR,USA;2012.p.997.
[50]Daley DJ,Kendall DG.Epidemics and rumors.Nature1964;204(4963):1118.
[51]Moreno Y,Nekovee M,Pacheco AF.Dynamics of rumor spreading in complex
networks.Phys Rev E Stat Nonlin Soft Matter Phys2004;69(6Pt2):066130.
[52]Nekovee M,Moreno Y,Bianconi G,Marsili M.Theory of rumor spreading in
complex social networks.Phys A2007;374(1):457–70.
[53]Zhou J,Liu Z,Li B.In?uence of network structure on rumor propagation.Phys
Lett A2007;368(6):458–63.
[54]Wang H,Deng L,Xie F,Xu H,Han J.A new rumor propagation model on
SNS structure.In:Proceedings of the2012IEEE International Conference on Granular Computing;2012Aug11–13;Hangzhou,China;2012.p.
499–503.
[55]Wang Y,Yang X,Han Y,Wang X.Rumor spreading model with trust mechanism
in complex social https://www.360docs.net/doc/664755022.html,mum Theor Phys2013;59(4):510–6.
[56]Xia L,Jiang G,Song B,Song Y.Rumor spreading model considering hesitating
mechanism in complex social networks.Phys A2015;437:295–303.
[57]Su Q,Huang J,Zhao X.An information propagation model considering
incomplete reading behavior in microblog.Phys A2015;419:55–63.
[58]Liu Q,Li T,Sun M.The analysis of an SEIR rumor propagation model on
heterogeneous network.Phys A2017;469:372–80.
[59]Zhao L,Wang J,Chen Y,Wang Q,Cheng J,Cui H.SIHR rumor spreading model
in social networks.Phys A2012;391(7):2444–53.
[60]Zhao L,Qiu X,Wang X,Wang J.Rumor spreading model considering forgetting
and remembering mechanisms in inhomogeneous networks.Phys A2013;392
(4):987–94.
[61]Leskovec J,Krause A,Guestrin C,Faloutsos C,VanBriesen J,Glance N.Cost-
effective outbreak detection in networks.In:Proceedings of the13th ACM SIGKDD International Conference on Knowledge Discovery and Data Mining;
2007Aug12–15;San Jose,CA,USA;2007.p.420–9.
[62]Zhou C,Zhang P,Zang W,Guo L.On the upper bounds of spread for greedy
algorithms in social network in?uence maximization.IEEE Trans Knowl Data Eng2015;27(10):2770–83.
[63]Chen W,Wang C,Wang Y.Scalable in?uence maximization for prevalent viral
marketing in large-scale social networks.In:Proceedings of the16th ACM SIGKDD International Conference on Knowledge Discovery and Data Mining;
2010Jul25–28,Washington,DC,USA;2010.p.1029–38.
[64]Goyal A,Lu W,Lakshmanan LVS.SIMPATH:An ef?cient algorithm for in?uence
maximization under the linear threshold model.In:Proceedings of the2011 IEEE11th International Conference on Data Mining;2011Dec11–14;
Vancouver,BC,Canada;2012.p.211–20.
[65]Jung K,Heo W,Chen W.IRIE:a scalable in?uence maximization algorithm
for independent cascade model and its extensions.Rev Crim2011;56
(10):1451–5.[66]Borgs C,Brautbar M,Chayes J,Lucier B.Maximizing social in?uence in nearly
optimal time.In:Proceedings of the25th Annual ACM-SIAM Symposium on Discrete Algorithms;2014Jan5–7;Portland,OR,USA;2014.p.946–57. [67]Tang Y,Xiao X,Shi Y.In?uence maximization:Near-optimal time complexity
meets practical ef?ciency.In:Proceedings of the2014ACM SIGMOD International Conference on Management of Data;2014June22–27;
Snowbird,UT,USA;2014.p.75–86.
[68]Li CT,Lin SD,Shan MK.In?uence propagation and maximization for
heterogeneous social networks.In:Proceedings of the21st International Conference on World Wide Web;2012Apr16–20;Lyon,France;2012.p.559–
60.
[69]Subbian K,Sharma D,Wen Z,Srivastava J.Social capital:The power of
in?uencers in networks.In:Proceedings of the2013International Conference on Autonomous Agents and Multi-agent Systems;2013May6–10;Saint Paul, MN,USA;2013.p.1243–4.
[70]Li H,Bhowmick SS,Sun A.CINEMA:Conformity-aware greedy algorithm for
in?uence maximization in online social networks.In:Proceedings of the16th International Conference on Extending Database Technology;2013Mar18–22;Genoa,Italy;2013.p.323–34.
[71]Lee JR,Chung CW.A query approach for in?uence maximization on speci?c
users in social networks.IEEE Trans Knowl Data Eng2015;27(2):340–53. [72]Deng X,Pan Y,Shen H,Gui J.Credit distribution for in?uence maximization in
online social networks with node features.J Intell Fuzzy Syst2016;31
(2):979–90.
[73]Yao Q,Zhou C,Shi R,Wang P,Guo L.Topic-aware social in?uence
minimization.In:Proceedings of the24th International Conference on World Wide Web;2015May18–22;Florence,Italy;2015.p.139–40.
[74]Wang B,Chen G,Fu L,Song L,Wang X.DRIMUX:dynamic rumor in?uence
minimization with user experience in social networks.IEEE Trans Knowl Data Eng2017;29(10):2168–81.
[75]Groeber P,Lorenz J,Schweitzer F.Dissonance minimization as a
microfoundation of social in?uence in models of opinion formation.J Math Sociol2014;38(3):147–74.
[76]Chang CW,Yeh MY,Chuang KT.On the guarantee of containment probability
in in?uence minimization.In:Proceedings of the2016IEEE/ACM International Conference on Advances in Social Networks Analysis and Mining;2016Aug 18–21;San Francisco,CA,USA;2016.p.231–8.
[77]Faisan MM,Bhavani SD.Maximizing information or in?uence spread using
?ow authority model in social networks.In:Proceedings of the10th International Conference on Distributed Computing and Internet Technology;2014Feb6–9;Bhubaneswar,India;2014.p.233–8.
[78]Subbian K,Aggarwal C,Srivastava J.Content-centric?ow mining for in?uence
analysis in social streams.In:Proceedings of the22nd ACM International Conference on Information&Knowledge Management;2013Oct27–Nov1;
San Francisco,CA,USA;2013.p.841–6.
[79]Kutzkov K,Bifet A,Bonchi F,Gionis A.STRIP:Stream learning of in?uence
probabilities.In:Proceedings of the19th ACM SIGKDD International Conference on Knowledge Discovery and Data Mining;2013Aug11–14;
Chicago,IL,USA;2013.p.275–83.
[80]Teng X,Pei S,Morone F,Makse HA.Collective in?uence of multiple spreaders
evaluated by tracing real information?ow in large-scale social networks.Sci Rep2016;6(1):36043.
[81]Chintakunta H,Gentimis A.In?uence of topology in information?ow in social
networks.In:Proceedings of the2016Asilomar Conference on Signals, Systems and Computers;2016Nov6–9;Paci?c Grove,CA,USA;2017.
p.67–71.
[82]Subbian K,Sharma D,Wen Z,Srivastava J.Finding in?uencers in networks
using social capital.In:Proceedings of the2013IEEE/ACM International Conference on Advances in Social Networks Analysis and Mining;2013Aug 25–28;Niagara Falls,ON,Canada;2013.p.592–9.
[83]Liu G,Zhu F,Zheng K,Liu A,Li Z,Zhao L,et al.TOSI:a trust-oriented social
in?uence evaluation method in contextual social networks.Neurocomputing 2016;210:130–40.
[84]Deng X,Pan Y,Wu Y,Gui J.Credit distribution and in?uence maximization in
online social networks using node features.In:Proceedings of the201512th International Conference on Fuzzy Systems and Knowledge Discovery;2015 Aug15–17;Zhangjiajie,China;2016.p.2093–100.
[85]Zhu Y,Wu W,Bi Y,Wu L,Jiang Y,Xu W.Better approximation algorithms for
in?uence maximization in online social networks.J Comb Optim2015;30
(1):97–108.
[86]He J,Hu M,Shi M,Liu Y.Research on the measure method of complaint theme
in?uence on online social network.Expert Syst Appl2014;41(13):6039–46.
[87]Gehrke J,Ginsparg P,Kleinberg J.Overview of the2003KDD cup.ACM SIGKDD
Explor Newslett2003;5(2):149–51.
[88]Yang J,Leskovec J.De?ning and evaluating network communities based on
ground-truth.Knowl Inf Syst2015;42(1):181–213.
[89]Tang J,Zhang J,Yao L,Li J,Zhang L,Su Z.ArnetMiner:Extraction and mining of
academic social networks.In:Proceedings of the14th ACM SIGKDD International Conference on Knowledge Discovery and Data Mining;2008 Aug24–27;Las Vegas,NV,USA;2008.p.990–8.
46K.Li et al./Engineering4(2018)40–46
Engineering 2 (2016) xxx–xxx
Research
Cybersecurity—Review
社会影响力分析——模型、方法和评价
李侃*, 张林, 黄河燕
School of Computer Science and Technology, Beijing Institute of Technology, Beijing 100081, China
a r t i c l e i n f o摘要
Article history:
Received 10 December 2017 Revised 5 January 2018 Accepted 8 January 2018 Available online 16 February 2018社会影响力分析（SIA）是一个广泛的研究领域，吸引了诸多研究者的兴趣。本文总结了SIA相关的模型、方法和评价方面代表性的工作，并将SIA模型归纳为两种类型：微观模型和宏观模型。微观模型考虑人与人之间的相互影响和影响的过程，而宏观模型认为每个人具有相同的传播概率和影响力。本文分析了包括影响最大化、影响最小化、影响流和个人影响力等的社会影响力分析方法；介绍了影响力评价指标，并分析了社会影响力评价模型。本文的目标是对社会影响力提供全面的分析，旨在辅助理解社会行为，为舆论影响提供理论基础，并揭示未来的研究方向和潜在的应用。? 2018 THE AUTHORS. Published by Elsevier LTD on behalf of Chinese Academy of Engineering and Higher Education Press Limited Company This is an open access article under the CC BY-NC-ND license
关键词
社会影响力分析
在线社交网络
社会影响力分析模型影响力评价
1.引言
在线社交网络，如微博、Twitter和Facebook等，为用户的信息传播提供了有价值的平台。在信息传播过程中，社会影响力伴随一个人的观点、情绪或行为受到其他人的影响而产生[1]。个人的态度、想法、感受或行为会因为与其他人或团体发生交流而改变。社会影响力分析（SIA）正在成为社交网络中一个重要的研究领域。SIA主要研究如何模拟网络上影响扩散的过程，以及如何提出一种有效的方法来识别网络中的相应的目标节点[2]。这些研究包括：谁影响了谁、谁受到了影响、谁是最有影响力的人等。SIA具有重要的社会意义，并已在许多领域得到了应用。病毒营销[3–10]、在线推荐[11]、医疗社区[12–14]、专家发现[15–17]、谣言扩散[18]和其他领域[19–21]都依赖于社会影响效应。分析社会影响力可以帮助理解人的社会行为，为公众决策和影响舆论提供理论支持，并可促进各类活动的交流和传播[22]。
本文从模型、方法、评价等方面对SIA进行了全面的分析。基于此，本文分析研究了现有模型、方法以及社会影响力评价模型的优缺点。首先介绍了现有的社会影响力分析模型；然后总结了社会影响力相关计算方法；最后分析了社会影响力评价的相关工作。
本文的组织如下：第2节讨论了SIA模型；第3节分析了包括影响最大化、影响最小化、影响流和个人影响力等的SIA方法；第4节详细论述了社会影响力评价工作；最后，本文总结了所提到的社会影响力模型和方法，并讨论了这个领域开放的问题。
* Corresponding author.
E-mail address: likan@https://www.360docs.net/doc/664755022.html, (K. Li) Contents lists available at ScienceDirect Engineering
46Author name et al. / Engineering 2(2016) xxx–xxx
2.社会影响力分析模型
SIA模型已经被广泛地研究，本文将这些模型归纳为两类：微观模型和宏观模型。
2.1. 微观模型
微观模型关注人与人之间的相互作用，并考虑影响过程。广泛使用的两个影响力分析模型是独立级联模型（independent cascade，IC）[23–25]和线性阈值模型（linear threshold，LT）[23,26]。这两种模型最早由Kempe等[23]提出，现在仍主要用于模拟社会影响力扩散过程。
2.1.1. IC和LT模型
IC模型将社会网络抽象为一个有向图G=(V, E)，其中V代表节点集合，E代表节点之间的有向边的集合，同时还存在一个种子节点集S(S?V)，S t(S t?V)表示时间t (t≥0)被激活的节点集合。在t+1时刻，每个节点v i会以概率P ij去激活它的非激活状态下的邻居节点。当没有节点被激活时，影响过程结束。在此过程中，每一个节点在被激活后有且只有一个机会去激活他的邻居节点，被激活节点会一直保持激活状态。
LT模型同样将网络抽象为一个有向图G=(V, E)，每一个节点与其所有邻居节点之间的影响权重之和满足∑v
j∈N g i act
w ij≤1。其中w ij表示节点v i和它的邻居节点v j之
间的影响权重；N g
i
act表示被节点v i激活的邻居节点集合；节点v i按照0和1之间的均匀分布随机选择自己的阈值θi。只有当节点v i激活状态下的邻居节点影响权重之和超过了v i的阈值，v i才会被激活。
2.1.2. 扩展的模型
在IC和LT模型中，为了计算某个节点的影响通常需要利用蒙特卡洛模拟过程，而这是一个非常耗时的操作，且不适合大规模社交网络。因此，很多研究者提出了IC和LT模型的改进模型。本文将这些改进模型划分为4类：基于IC的模型、基于LT的模型、基于IC和LT的模型、其他模型。这里只列举出在这些方面有代表性的模型。
（1）基于IC的模型。一些研究人员在IC模型基础之上，考虑了时间延迟和受时间限制的影响扩散模型。Chen等[27]扩展了IC模型并提出了具有相遇事件的IC 模型（IC model with meeting events ，IC-M）。在IC-M
模型中，激活节点以一定概率遇到非激活节点。虽然IC-M模型计算的时间复杂度略高于一般IC模型，但是计算结果更接近实际情况。Feng等 [28]基于以前的研究，发现信息重复曝光会减小对用户的影响。因此，他们在IC模型中加入了时间衰减因素，提出了一种基于传播路径的算法来评价种子节点的影响力。在提出的新模型中社会网络的每个有向边有两个值：传播概率和预计传播延迟时间。Mohamadi-Baghmolaei等[29]考虑了时间和信任因素，提出了一种基于信任的延迟感知独立级联模型（trust-based latency-aware independent cascade，TLIC）。这是首次在经典IC模型中考虑信任因素。在TLIC模型中，信任邻居节点和不信任邻居节点按照不同的概率改变一个节点的状态（如激活状态或者非激活状态）。Budak等[30]研究了在社会网络中两种影响相互竞争和相互影响抵消的过程，提出了多层IC模型，该模型可以同时模拟两个级联传播的过程。
（2）基于LT的模型。Liu等[31]从遏制社会网络中竞争对手的影响扩散出发，将LT模型改进为扩散-遏制模型（diffusion-containment，D-C）。传统的LT模式不适用于涉及扩散和遏制影响的情况。在D-C模型中，某个节点的状态受激活概率的控制，每个节点只会受到传播概率较高的邻居节点的影响。同样每个节点和所有邻居节点之间的传播概率之和小于或等于1。Borodin等[32]分析了基于LT模型的不同模型的竞争影响扩散效果。
（3）基于IC和LT的模型。Mohammadi等[33]考虑了时间延迟提出了两种扩散模型：延迟独立级联模型（de-layed independent cascade，DIC）和延迟线性阈值模型（delayed linear threshold，DLT）。在这两个模型中，节点有3种状态：活动、非活动和潜在活动。为了从非活动状态转变为活动状态，节点必须经过处于潜在活动状态的中间过程。与传统的IC和LT模型相比，该模型中影响扩散的效果更好。但由于存在多个状态，因此计算复杂度相对较高。在IC和LT模型中，信息扩散被视为一系列以同步方式发生的节点状态变化。但是，实际的扩散是以异步方式进行的，并且已知数据的时间戳并不是均匀分布。因此有些模型放弃了传统IC和LT模型的同步假设，并扩展了这两种模型以使状态可以异步变化。Sait等[34]对IC和LT模型进行了异步扩展，提出了异步独立级联模型（asynchronous IC，AsIC）和异步线性阈值模型（asynchronous LT，AsLT）。他们的算法可以从相对较少的信息扩散结果中估计网络中节点的影响程
47 Author name et al. / Engineering 2(2016) xxx–xxx
度，从而避免过度拟合问题。Guille和Hacid[35]也提出了一个异步模型来模拟扩散过程——基于时间的异步独立级联模型（time-based AsIC）。其他研究者[30,32,36–44]也在不同的方向对经典模型进行了改进。Fa等[38]构建了两个模型：机会主义一对一模型（the opportunistic one-activate-one，OPOAO）和确定性一对多模型（the deterministic one-activate-many，DOAM）。在OPOAO模型中每个人只能同时与另外一个人交流，而DOAM模型类似于信息广播过程的机制。对于OPOAO模型，他们使用经典的贪婪算法产生一个(1–1/e)的近似解；对于DOAM模型，他们提出了一种基于覆盖的贪婪算法（set cover-based greedy，SCBG）来实现O(ln n)求解（n是节点数），这两个模型中每个用户的传播概率都是相同的。Galam[45]提出了一个调查意见的动态模型，Lee等[39]对Galam提出的模型进行改进，但是这些模型仅限于口碑信息传播的过程。
（4）其他模型。一些模型与IC或LT模型及其变体不同，它们从新的角度解决了信息影响扩散问题。Lee等[46]提出了一个数据驱动的模型，利用元学习概念求解影响力最大化。但这种模式需要大量的数据，其结果的准确性还需要进一步改善。Golnari等[47]提出了一个热传导模型（heat conduction，HC）。与之前仅考虑逐步传播过程的IC或LT模型完全不同，HC模型考虑非逐步传播过程，传播初始化时有一组种子节点，其他节点可以任意赋值。Wang等[48]研究了在大规模图像社交网络中的情感影响，并提出了情感影响模型。他们设计了一个因子图模型来推断来自社交网络中图像的情感影响。Gao[49]提出了一个读写模型（read-write，RW）来描述意见形成、影响和扩散的详细过程。然而，这个模型需要进一步考虑3个主要问题：许多模型必需的参数怎样推断、有关意见影响和扩散的数据集怎样收集以及适合于此任务的评价指标。
无论是经典IC、LT模型或新模型，社会影响力分析模型的最终目标都是利用更少的内存实现更快的计算和更精确的结果，但是这些微观模型缺乏一些细节。首先，由于个体特征的差异（即教育背景和个人意识），不同的人对同一信息的接受程度不同；其次，不同的人对他人的影响力不同，即信息发布者的权威不同对邻居和社会的影响力也不同；第三，个人将信息传递给其他人的概率并不是恒定的，应该取决于个人对信息的感兴趣程度。而且这些模型都使用二元变量来记录个体是否受到影响，假设一旦一个人受到影响，这个人再也不会改变其状态。然而，这种假设并不能真实反映出一个人从一种状态向另一种状态的自然变化。
2.2. 宏观模型
宏观模型认为所有人对信息具有相同的兴趣，同时也具有相同的传播概率和相同的影响力。由于宏观模型不考虑个体，所以这些模型计算结果的准确率较低。因此，为了改进这些模型，应该考虑个体之间的差异。宏观模型将节点划分为不同的类别（即状态），并关注每个类别中节点的状态演变。每个类别中节点的百分比用简单的微分方程表示。传染病模型是用于从宏观角度研究社会影响的最常见模型，这些模型主要是为了模拟传染病传播过程，但是它们忽视了社交网络的拓扑特征。每个类别中节点的百分比的计算方式过于简单，不能准确描述这种复杂的演化过程。Daley和Kendall[50]分析了传染病扩散与信息传播之间的相似性，并提出了经典的Daley-Kendall模型。从那以后，研究人员不断改进这些传染病模型来克服它们的缺陷。Daley和Kendall[50]、Moreno等[51]、Nekovee等[52]和Zhou等[53]在他们的方法中考虑网络底层的拓扑特征，然而忽视了影响扩散过程中个人行为的影响。
最近，研究人员开始考虑信息影响扩散过程中人类行为和不同机制的作用。Zhao等[59,60]在同质和非同质网络中提出了susceptible-infected-hibernator-removed模型作为经典susceptible-infected-removed模型（SIR）的延伸，以便纳入遗忘和记忆机制。Wang等[54]在概率模型中考虑了社会的正面或负面影响，分析了社会强化如何影响传播速度提出了一个在线社交网络中的谣言扩散模型的变体。Wan等[55]通过引入节点之间的信任机制提出了一个SIR模型，这种机制可以减少谣言最终影响范围和谣言扩散速度。Xia等[56]考虑到谣言内容的吸引力和模糊性，并且认为谣言越清晰，其影响越小，提出了一个修改后的susceptible-exposed-infected-removed模型（SEIR）用于研究犹豫机制对谣言传播模型的影响。Su等[57]通过考虑用户的不完整阅读行为，提出了microblog-susceptible- infected-removed的信息传播模型。Liu等[58]受到Xia等[56]和Su等[57]工作的激发，在SEIR模型基础上提出了一个新的异构网络SEIR模型用以研究微博的扩散动态。3. 社会影响力分析方法
社会影响力分析方法用于解决影响最大化、影响最
48Author name et al. / Engineering 2(2016) xxx–xxx
小化、影响传播和个体影响力等社会网络影响分析的子问题。所有这些问题都涉及影响的扩散，所以在某些情况下，同一个影响传播模型可以用于所有这些问题。但是对于每一个问题，使用该模型的最终目的不同。
3.1. 影响最大化
影响最大化需要找到社交网络中最具影响力的成员集合。Kempe等[23]对这个问题进行了定义：给定以用户为节点的有向图以及预算或阈值数k，其中有向图中边的权重代表用户之间相互影响的程度，影响最大化的目的是在社交网络中找到k个节点，以便通过激活这些节点得到最大的影响传播范围。这是一个离散优化问题，同时，对于IC和LT模型都是NP难问题。影响最大化是社会影响力分析中研究最为广泛的问题，Leskovec等[10]和Rogers [20]将其当做一个算法问题，并提出了一些概率计算方法。影响最大化算法需要满足计算速度快、结果精度高和计算存储要求低这3个标准。考虑个体差异的大多数算法可以分为贪婪算法或启发式算法。
3.1.1. 贪婪算法
贪婪算法在每次迭代中“贪婪地”选择对现有种子集具有最大边际增益的活动节点。贪婪算法的研究是基于爬山贪婪算法，在这种算法中每次选择都可以利用局部最优解来逼近全局最优解，把影响力最大的节点加入种子集合中。社会影响力分析中很多用于研究个体差异的算法都是基于这些贪婪算法，这些算法的优点是精度比较高，达到了(1–1/e–ε)，ε可以是任何大于0的数值，然而算法复杂度高、执行时间长、效率相对较低。Kempe等[23]首次把求解影响最大化作为离散优化问题，并提出了一种贪婪近似算法。Leskovec等[61]提出了一种贪婪的优化方法——cost-effective lazy forward （CELF）方法。Chen等[27]提出了新的贪婪算法：New-Greedy算法和MixGreedy算法。Zhou等[62]提出了基于上限的延迟转发算法（upper bound-based lazy forward，UBLF），用来发现最具影响力的k个节点。
3.1.2. 启发式算法
由于贪婪算法的计算复杂度很高，许多优秀的启发式算法已经被提出用来缩短求解时间并追求更高的算法效率。这些启发式算法基于特定启发信息（如节点的度或PageRank排序结果）迭代地选择节点，而不用每次迭代过程中计算节点的边际增益。但是启发式算法的缺点是准确性相对较低。最基本的启发式算法是由Kempe 等[23]提出的随机算法、基于度的算法和中心启发式算法。Chen等[41]根据基于度的启发式算法提出了独立级联模型（IC）的启发式算法——DegreeDiscount，之后他们还提出了一种新的启发式算法——prefix excluding maximum influence arborescence（PMIA）算法[63]。对于LT模型，Chen等[63]提出了基于局部有向无环图启发式算法（local directed acyclic graph，LDAG）。还有很多启发式算法也是基于这个启发式思想，如SIM-PATH[64]和IRIE[65]。除此之外，Borgs等[66]在IC模型下提出了一个线性时间复杂度用于求解影响最大化的算法，取得了理论突破。Tang等[67]提出了一个两阶段影响最大化的两阶段算法（two-phase influence maximization，TIM），这个算法的时间复杂度是O(k+l) (n+m)log n/ε2，求解结果以至少(1–n–l)的概率符合(1–1/ e–ε)近似。
上述算法的时间复杂度如表1所示。
在这里，本文介绍了一些考虑到个体差异或用户自身属性的代表性研究。Li等[68]为了模拟异构社交网络对解决影响最大化问题的作用，考虑了人类之间的行为关系（即正常和稀有行为）。他们提出了两种基于熵的启发式算法来识别网络中的传播者，然后最大化传播的
表1 不同算法时间复杂度比较
Algorithm Time complexity
Hill-climbing greedy[23]O(knRm)
CELF[61]O(knRm)
NewGreedyIC[41]O(kRm)
NewGreedyWC[41]O(kRTm)
MixGreedyIC[41]O(kRm)
MixGreedyWC[41]O(kRTm) DegreeDiscountIC[41]O(k log n + m)
PMIA[63]O[nt iθ + kn oθn iθ(n iθ + log n) LDAG[63]O(l v + n log l v)
SIMPATH[64]O(kmn)
IRIE[65]O(kmn)
Quasilinear time algorithm[66]O[kε–2(m + n) log n]
TIM[67]O[(k + l)(m + n) log n/ε2]
n: number of vertices in G; m: number of edges in G; k: number of seeds to be selected; R: number of rounds of simulations; T: number of iterations; t iθ, n oθ, n iθ: constants decided by θ; θ: the influence threshold; n iθ = max v∈V {|MIIA(v; θ)|}; n oθ = max v∈V {|MI O A(v; θ)|}; MIIA(v; θ)/MI O A(v; θ): the maximum influence in arborescence/out arborescence of a node v; t iθ: the maximum run-ning time to compute MIIA(v; θ); l: number of communities in the network; ε: any constant larger than 0; l v: the volume of LDAG(v; θ).
49 Author name et al. / Engineering 2(2016) xxx–xxx
影响。尽管基于熵的启发式性能优于Degree [23]和De-greeDiscount[41]，但该方法仍然基于启发式思想，准确性不足。Subbian等[69]提出个人社会价值的概念，他们发现现有的影响最大化算法不能模拟个人的社会价值，但是这通常是相互联系的真正动机。因此他们基于社会资本的概念提出了一个框架，通过网络中的绑定和桥接连接的数量来计算网络的社会价值。该算法的性能优于PMIA[63]，PageRank和基于度的方法。Li等[70]提出了从众意识级联模型（conformity-aware cascade，C2）和从众感知贪婪算法来解决影响最大化问题。这个算法适用于分布式平台时效果很好，但由于基于贪婪算法并考虑从众意识级联，所以计算的复杂性仍需要改善。Lee 和Chung[71]将影响最大化问题定义为查询处理问题，即区分特定用户与其他用户。由于影响最大化查询处理是NP难问题，且求解目标函数也是NP难问题，所以他们侧重如何有效地近似求解最优种子节点。Deng等[72]发现求解影响最大化问题中在计算不同用户的影响时，节点特征总是被忽视。因此他们提出了3种量化措施来分别评价节点特征：用户活跃度、用户敏感度和用户亲和力。他们将节点特征融合到用户的静态影响中，然后使用连续指数衰减函数来转换两个相邻用户之间用户动态影响的强度。他们还提出了带有节点特征的信用分布模型（credit distribution with node feature，CD-NF），该模型重新定义了信用，并且基于CD-NF模型设计了具有节点特征的贪婪算法（greedy algorithm with node fea-ture，GNF）。
考虑影响最大化问题时，个人的属性会对结果产生很大的影响；因此，需要考虑到个体差异或用户的属性。从以上关于个体差异研究的描述中可以清楚地看出，大多数研究都是基于贪婪算法以追求高准确度，然而由于考虑个体差异，与原始贪婪算法计算相比，计算复杂度将进一步提高，导致算法效率低。因此，目前需要改善计算的复杂度和结果的准确率，为了缩短运行时间、提高算法效率，可以考虑使用启发式算法，也可以在分布式平台上进行大量工作。
3.2. 影响最小化
假设负面信息在网络G=(V, E)中传播，最初被影响的节点集合为S(S?V)，影响最小化的目标是指通过干预阻塞k个节点的传播使最终受影响的范围最小。阻塞的节点集合用D表示D?V，而且k (<<|V|)是一个常数，可以表示为以下优化问题：
（1）式中，σ(S|V\D)表示当集合D中的节点被阻塞后，S集中节点的影响力。
从这个概念[73]中，我们知道影响最小化是影响最大化的对偶问题。影响最小化主要用于遏制谣言、监督舆论等。Yao等[73]提出了一种方法，通过从主题模型的角度阻止有限数量的节点传播来减少网络中的负面影响的传播。当社交网络中出现谣言和其他不良事件时，已有一些用户受影响，这个模型的目的是最大限度地减少最终受到影响的用户数量。Wang等[74]提出了一个考虑用户感受的动态谣言影响最小化模型。这个模型通过阻止部分节点激活，最大限度地减少了谣言的影响。Groeber等[75]受社会心理学中认知失调的概念启发设计了一个社会影响力的总体框架，其中个体将因与邻居在某一社交网络上的不一致而产生的不和谐的先决条件最小化。Chang等[76]提出了一种用于估计成功限制受影响数量低于失控阈值的概率，它可以允许当局适应性地调整干预成本以满足必要的风险控制，避免爆发风险。他们还提出了一个影响最小化模型，以有效防止网络上易发生大规模舆情的扩散。
3.3. 影响流
信息传播会伴随着影响流[77]。近年来，许多研究人员提出了研究影响流的方法。Subbian等[78]提出了一种具有特定流动有效性约束条件的流模式挖掘方法。Kutzkov等[79]提出了一种叫STRIP的流式方法来计算社交网络中每个链接的影响力。Teng等[80]研究了包括美国物理学会、Facebook、Twitter和LiveJournal在内的各种真实平台的信息流，然后利用用户的行为模式构建虚拟信息影响扩散过程。Chintakunta和Gentimis[81]讨论了社交网络的拓扑结构与其内部信息流之间的关系。然而，大多数社交网络与微博平台不同，不能提供足够的上下文来挖掘流模式。
3.4. 个体影响力
个体影响力是一种相对微观的评估，它可以模拟用户对其他用户或整个社交网络的影响。Chintakunta和Gentimis[82]提出了一种名为SoCap的方法在社交网络中找到有影响力的人，他们将在社交网络中的发现有影响力的人建模为价值分配问题，其中分配的价值代表个人社会资本。Subbian等[82]提出了一种在开放式智
50Author name et al. / Engineering 2(2016) xxx–xxx
能体系统中，使用矩阵分解方法，通过度量网络中节点的影响力来挖掘有影响的智能体的方法。Liu等[83]提出了面向信任的社会影响力方法（trust-oriented social influence，TOSI），该方法考虑了社会背景（即参与者之间的社会关系和社会信任）和用户偏好，来评价个人影响力。TOSI方法在有效性、效率和鲁棒性方面大大优于SoCap。Deng等[84]在评估不同用户的影响时，纳入了时间因素和节点的特征。结果表明，他们的方法对于识别种子节点是合理有效的，并且其对影响扩散的预测结果比原来忽略了扩散过程中节点特征的方法更准确。
总之，考虑更全面的用户特征和用户交互信息会带来更精确的结果。
4. 社会影响力分析评价
4.1. 分析评价指标
运行时间是一个评价模型效率非常直观的度量标准，而且易于计算。一般来说，在相同条件下，模型运行得越快，效果就越好。然而，传统的贪婪算法需要通过大量重复的蒙特卡洛模拟来计算节点的影响力，导致消耗很长的时间，特别是在当前大规模社交网络上，现有的算法无法满足应用对效率的要求。因此，运行时间是衡量的重要手段。
由于求解影响传播问题是NP难问题，因此很难获得目标函数的最优解。现在的大多数算法都依赖函数的单调性和子模性来满足(1–1/e)的近似解[23]。然而，尝试实现更高的近似求解从未停止过。Zhu等[85]考虑影响传递性和限制传播距离，在他们的模型中提出了基于半定的算法。
另一个指标是蒙特卡罗模拟的次数。由于无法获得最优解，因此通常使用蒙特卡洛模拟来估计实际价值。现有的贪婪算法需要在初始阶段对每个节点的传播函数进行重复的蒙特卡洛模拟计算，这大大降低了模型的效率。Zhou等[62]提出的UBLF算法可以减少CELF方法中蒙特卡罗模拟数量的95％以上，并且当种子集小时，可以实现了2~10倍的加速。
预期传播程度表示种子集最终可能影响的节点数量，受影响的节点数量越多越好。在现实生活场景中有许多需要尽可能最大影响范围的应用，这种应用的典型例子是市场营销和广告。在这两种应用中，最终的预期结果为产品促销的好处或产品的盈利能力。因此，探索高预期的传播算法是社会影响力分析的一个重要研究内容。
鲁棒性是指在维持一些其他性能不变的前提下，允许某些参数（如结构、大小）扰动的特性。Jung等[65]和Liu等[83]在他们的算法中都提到了鲁棒性。Jung等[65]提出的IRIE算法在各种密度网络和不同大小级联下运行时间和内存使用方面更稳健和稳定。实验结果表明，大规模网络中IRIE算法的运行速度比PMIA[63]等现有方法运行速度快两个数量级，并且使用的内存更少。Liu等[83]提出的TOSI评估方法稳定性比SoCap[81]表现得更出色。
可伸缩性是在对现有系统影响很小的情况下能够不断扩展或增强系统的功能，在社会网络中可伸缩性通常是指从小型网络扩展到大型网络的能力，这是用于评估模型质量的常用指标。由于算法复杂度高、运行时间长，目前的算法仅适用于节点数在100万以下的中小型社会网络，鉴于当今的大型社交网络，必须设计具有良好可扩展性的影响力分析算法，以应对大量社会网络数据带来的挑战。
4.2. 社会影响力评价模型
社会影响力评价是一个复杂的过程。社会关系作为一种主观属性，具有动态性、事件差异性、不对称性、传递性等多种特征，而且在社会网络中频繁的用户互动和网络结构的变化使得社会影响力评价更加困难。关于社会影响力评价模型的研究一直都有学者在研究。He等[86]设计了一个关于在线投诉主题的影响度量模型，该模型基于熵权模型、实时监控和分析投诉信息的静态和动态属性。企业可以使用这种模式来管理在线用户的投诉。Wang等[12]提出了一个细粒度的基于特征的社会影响力评价模型（fine-grained feature-based social influence，FBI），该模型探讨了用户的重要性以及用户影响他人的可能性。然后他们通过对朋友的影响力贡献设计一个基于PageRank算法的社会影响力评价模型。FBI评价模型在HEPTH[87]、DBLP[88]和Arnet-Miner[89]3个数据集上验证可以出色识别出影响较大的TOP-K用户。
5. 结论和下一步工作
本文从模型、方法和评价等方面综述了SIA的最新研究进展，分析了当前模型和方法的优缺点，并揭示了未来的研究方向和潜在的应用。
在社会影响力分析模型中，本文归纳了两种类型的
51 Author name et al. / Engineering 2(2016) xxx–xxx
模型：微观模型和宏观模型。微观模型考虑人的相互作用和影响过程；宏观模型认为所有用户有相同的传播概率和相同的影响力，未来的宏观模型应着眼于信息传播过程中如何加入人的行为和实现不同的机制。虽然许多研究者在改进经典模型和从不同视角提出新模型方面做了大量的努力，例如在模型中加入约束和引入竞争影响传播，但仍有改进的余地。
在大多数的现有模型中，考虑的都是单一网络中一个人受其他人的影响，而且影响扩散过程是独立的，然而在现实生活中，人们经常在多网络中与他人交流。在这种情况下，社会影响发生在多网络，不同的单一网络之间的影响会存在合作和竞争关系。如何在多网络中构建信息传播模型是一个有价值的研究课题。同时在动态网络中如何计算跨时间影响传播的问题也值得被研究。除此之外，在大多数实验中数据集涵盖了大约100 000个节点，因此将社交网络分析相关问题如何应用于海量数据集（可能包括数百万或数千万甚至更多个节点）仍需要研究。简而言之，在扩展SIA模型以解决效率和可伸缩性等局限性的问题上仍有研究空间。
致谢
本文研究部分得到了中国国家重点基础研究项目（2013CB329605）的支持。感谢李玲玲和商军英对本文的建设性意见和支持。Compliance with ethics guidelines
Kan Li, Lin Zhang and Heyan Huang declare that they have no conflict of interest or financial conflicts to disclose. References
[1] Travers J, Milgram S. The small world problem. Psychol Today 1967;1:61–7.
[2] Chen W, Lakshmanan LV, Castillo C. Information and influence propagation in
social networks. San Rafael: Morgan & Claypool; 2013.
[3] Freeman LC. A set of measures of centrality based on betweenness. Sociome-
try 1977;40(1):35–41.
[4] Baas F. A new product growth model for consumer durables. Manage Sci
1969;15(5):215–27.
[5] Brown JJ, Reingen PH. Social ties and word-of-mouth referral behavior. J Con-
sum Res 1987;14(3):350–62.
[6] Mahajan V, Muller E, Bass FM. New product diffusion models in marketing: A
review and directions for research. J Mark 1990;54(1):1–26.
[7] Domingos P, Richardson M. Mining the network value of customers. In: Pro-
ceedings of the 7th ACM SIGKDD International Conference on Knowledge Discovery and Data Mining; 2001 Aug 26–29; San Francisco, CA, USA; 2001. p.
57–66.
[8] Goldenberg J, Libai B, Muller E. Talk of the network: a complex systems look
at the underlying process of word-of-mouth. Mark Lett 2001;12(3):211–23.
[9] Richardson M, Domingos P. Mining knowledge-sharing sites for viral mar-
keting. In: Proceedings of the 8th ACM SIGKDD International Conference on
Knowledge Discovery and Data Mining; 2002 Jul 23–26; Edmonton, AB, Can-ada; 2002. p. 61–70.
[10] Leskovec J, Adamic LA, Huberman BA. The dynamics of viral marketing. J ACM
Trans Web 2007;1(1):5.
[11] Pálovics R, Benczúr AA, Kocsis L, Kiss T, Frigó E. Exploiting temporal influence
in online recommendation. In: Proceedings of the 8th ACM Conference on Recommender Systems; 2014 Oct 6–10; Foster City, CA, USA; 2014. p. 273–80.
[12] Wang G, Jiang W, Wu J, Xiong Z. Fine-grained feature-based social influence
evaluation in online social networks. IEEE Trans Parallel Distrib Syst 2014;25
(9):2286–96.
[13] Christakis NA, Fowler JH. The spread of obesity in a large social network over
32 years. N Engl J Med 2007;357(4):370–9.
[14] Fowler JH, Christakis NA. Dynamic spread of happiness in a large social net-
work: longitudinal analysis over 20 years in the Framingham Heart Study. Br Med J 2009;338(7685):23–7.
[15] Franks H, Griffiths N, Anand SS. Learning influence in complex social net-
works. In: Proceedings of the 2013 International Conference on Autonomous Agents and Multi-agent Systems; 2013 May 6–10; Saint Paul, MN, USA; 2013.
p. 447–54.
[16] Dong W, Pentland A. Modeling influence between experts. In: Proceedings
of the ICMI 2006 and IJCAI 2007 International Conference on Artificial In-telligence for Human Computing; 2006 Nov 3; Banff, AB, Canada; 2007. p.
170–89.
[17] Tang J, Sun J, Wang C, Yang Z. Social influence analysis in large-scale net-
works. In: Proceedings of the 15th ACM SIGKDD International Conference on Knowledge Discovery and Data Mining; 2009 Jun 28–Jul 1; Paris, France;
2009. p. 807–16.
[18] He Z, Cai Z, Wang X. Modeling propagation dynamics and developing opti-
mized countermeasures for rumor spreading in online social networks. In: Proceedings of the 2005 IEEE 35th International Conference on Distributed Computing Systems; 2015 Jun 29–Jul 2; Columbus, OH, USA; 2015. p. 205–14.
[19] Katz E, Lazarsfeld PF. Personal influence: The part played by people in the
flow of mass communications. New York: Free Press; 1965.
[20] Rogers EM. Diffusion of innovations. 5th ed. New York: Free Press; 2003.
[21] Keller E, Berry J. The influentials: one American in ten tells the other nine
how to vote, where to eat, and what to buy. New York: Free Press; 2003. [22] Peng S, Yang A, Cao L, Yu S, Xie D. Social influence modeling using informa-
tion theory in mobile social networks. Inf Sci 2017;379:146–59.
[23] Kempe D, Kleinberg J, Tardos é. Maximizing the spread of influence through a
social network. In: Proceedings of the 9th ACM SIGKDD International Confer-ence on Knowledge Discovery and Data Mining; 2003 Aug 24–27; Washing-ton, DC, USA; 2003. p. 137–46.
[24] Leskovec J, Mcglohon M, Faloutsos C, Glance NS, Hurst M. Patterns of cascad-
ing behavior in large blog graphs. In: Proceedings of the 2007 SIAM Interna-tional Conference on Data Mining; 2007 Apr 26–28; Minneapolis, MN, USA;
2007.
[25] Gruhl D, Guha R, Liben-Nowell D, Tomkins A. Information diffusion through
blogspace. In: Proceedings of the 13th International Conference on World Wide Web; 2004 May 17–20; New York, NY, USA; 2004. p. 491–501.
[26] Granovetter M. Threshold models of collective behavior. Am J Sociol 1978;83
(6):1420–43.
[27] Chen W, Lu W, Zhang N. Time-critical influence maximization in social net-
works with time-delayed diffusion process. In: Proceedings of the 26th AAAI Conference on Artificial Intelligence; 2012 Jul 22–26; Toronto, ON, Canada;
2012. p. 592–8.
[28] Feng S, Chen X, Cong G, Zeng Y, Chee YM, Xiang Y. Influence maximization
with novelty decay in social networks. In: Proceedings of the 28th AAAI Con-ference on Artificial Intelligence; 2014 Jul 27–31; Québec City, QC, Canada;
2014. p. 37–43.
[29] Mohamadi-Baghmolaei R, Mozafari N, Hamzeh A. Trust based latency aware
influence maximization in social networks. J Eng App Artif Intell 2015;41
(C):195–206.
[30] Budak C, Agrawal D, Abbadi AE. Limiting the spread of misinformation in so-
cial networks. In: Proceedings of the 20th International Conference on World Wide Web; 2011 Mar 28–Apr 1; Hyderabad, India; 2011. p. 665–74.
[31] Liu W, Yue K, Wu H, Li J, Liu D, Tang D. Containment of competitive influence
spread in social networks. Knowl Base Syst 2016;109(C):266–75.
[32] Borodin A, Filmus Y, Oren J. Threshold models for competitive influence in
social networks. In: Proceedings of the 6th International Conference on In-ternet and Network Economics; 2010 Dec 13–17; Stanford, CA, USA; 2010. p.
539–50.
[33] Mohammadi A, Saraee M, Mirzaei A. Time-sensitive influence maximization
in social networks. J Inf Sci 2015;41(6):765–78.
[34] Saito K, Ohara K, Yamagishi Y, Kimura M, Motoda H. Learning diffusion proba-
bility based on node attributes in social networks. In: Proceedings of the 19th International Conference on Foundations of Intelligent Systems; 2011 Jun 28–30; Warsaw, Poland; 2011. p. 153–62.
[35] Guille A, Hacid H. A predictive model for the temporal dynamics of informa-
tion diffusion in online social networks. In: Proceedings of the 21st Interna-tional Conference on World Wide Web; 2012 Apr 16–20; Lyon, France; 2012. p.
1145–52.
[36] Bharathi S, Kempe D, Salek M. Competitive influence maximization in social
networks. In: Proceedings of the 3rd International Conference on Internet and Network Economics; 2007 Dec 12–14; San Diego, CA, USA; 2007. p. 306–11. [37] Carnes T, Nagarajan C, Wild SM, Zuylen AV. Maximizing influence in a com-
52Author name et al. / Engineering 2(2016) xxx–xxx
petitive social network: A follower’s perspective. In: Proceedings of the 9th International Conference on Electronic Commerce; 2007 Aug 19–22; Minne-apolis, MN, USA; 2007. p. 351–60.
[38] Fan L, Lu Z, Wu W, Thuraisingham B, Ma H, Bi Y. Least cost rumor blocking
in social networks. In: Proceedings of the 2013 IEEE 33rd International Con-ference on Distributed Computing Systems; 2013 Jul 8–11; Philadelphia, PA, USA; 2013. p. 540–9.
[39] Lee W, Kim J, Yu H. CT-IC: Continuously activated and time-restricted inde-
pendent cascade model for viral marketing. In: Proceedings of the 2012 IEEE 12th International Conference on Data Mining; 2012 Dec 10–13; Brussels, Belgium; 2013. p. 960–5.
[40] Kostka J, Oswald YA, Wattenhofer R. Word of mouth: Rumor dissemination
in social networks. In: Proceedings of the 15th International Colloquium on Structural Information and Communication Complexity; 2008 Jun 17–20; Vil-lars-sur-Ollon, Switzerland; 2008. p. 185–96.
[41] Chen W,WangY, Yang S. Efficient influence maximization in social networks.
In: Proceedings of the 15th ACM SIGKDD International Conference on Knowl-edge Discovery and Data Mining; 2009 Jun 28–Jul 1; Paris, France; 2009. p.
199–208.
[42] Wang Y, Wang H, Li J, Gao H. Efficient influence maximization in weighted
independent cascade model. In: Proceedings of the 21st International Con-ference on Database Systems for Advanced Applications; 2016 Mar 27– 30;
Dallas, TX, USA; 2016. p. 49–64.
[43] Pathak N, Banerjee A, Srivastava J. A generalized linear threshold model for
multiple cascades. In: Proceedings of the 2010 IEEE International Conference on Data Mining; 2010 Dec 13–17; Sydney, Australia; 2010. p. 965–70.
[44] Bharathi S, Kempe D, Salek M. Competitive influence maximization in social
networks. In: Proceedings of the 3rd International Workshop on Web and Internet Economics; 2007 Dec 12–14; San Diego, CA, USA; 2007. p. 306–11. [45] Galam S. Modelling rumors: the no plane Pentagon French hoax case. Phys A
2003;320:571–80.
[46] Lin SC, Lin SD, Chen MS. A learning-based framework to handle multi-round
multi-party influence maximization on social networks. In: Proceedings of the 21st ACM SIGKDD International Conference on Knowledge Discovery and Data Mining; 2015 Aug 10–13; Sydney, Australia; 2015. p. 695–704.
[47] Golnari G, Asiaee A, Banerjee A, Zhang ZL. Revisiting non-progressive influ-
ence models: Scalable influence maximization. In: Proceedings of the 31st Conference on Uncertainty in Artificial Intelligence; 2015 Jul 12–16; Amster-dam, The Netherlands; 2015.
[48] Wang X, Jia J, Tang J, Wu B, Cai L, Xie L. Modeling emotion influence in image
social networks. IEEE Trans Affect Comp 2015;6(3):286–97.
[49] Gao D. Opinion influence and diffusion in social network. In: Proceedings of
the 35th International ACM SIGIR Conference on Research and Development in Information; 2012 Aug 12–16; Portland, OR, USA; 2012. p. 997.
[50] Daley DJ, Kendall DG. Epidemics and rumors. Nature 1964;204(4963):1118.
[51] Moreno Y, Nekovee M, Pacheco AF. Dynamics of rumor spreading in complex
networks. Phys Rev E Stat Nonlin Soft Matter Phys 2004;69(6 Pt 2):066130. [52] Nekovee M, Moreno Y, Bianconi G, Marsili M. Theory of rumor spreading in
complex social networks. Phys A 2007;374(1):457–70.
[53] Zhou J, Liu Z, Li B. Influence of network structure on rumor propagation. Phys
Lett A 2007;368(6):458–63.
[54] Wang H, Deng L, Xie F, Xu H, Han J. A new rumor propagation model on SNS
structure. In: Proceedings of the 2012 IEEE International Conference on Gran-ular Computing; 2012 Aug 11–13; Hangzhou, China; 2012. p. 499–503. [55] Wang Y, Yang X, Han Y, Wang X. Rumor spreading model with trust mecha-
nism in complex social networks. Commum Theor Phys 2013;59(4):510–6. [56] Xia L, Jiang G, Song B, Song Y. Rumor spreading model considering hesitating
mechanism in complex social networks. Phys A 2015;437:295–303.
[57] Su Q, Huang J, Zhao X. An information propagation model considering incom-
plete reading behavior in microblog. Phys A 2015;419:55–63.
[58] Liu Q, Li T, Sun M. The analysis of an SEIR rumor propagation model on heter-
ogeneous network. Phys A 2017;469:372–80.
[59] Zhao L, Wang J, Chen Y, Wang Q, Cheng J, Cui H. SIHR rumor spreading model
in social networks. Phys A 2012;391(7):2444–53.
[60] Zhao L, Qiu X, Wang X, Wang J. Rumor spreading model considering forget-
ting and remembering mechanisms in inhomogeneous networks. Phys A 2013;392 (4):987–94.
[61] Leskovec J, Krause A, Guestrin C, Faloutsos C, VanBriesen J, Glance N. Coste-
ffective outbreak detection in networks. In: Proceedings of the 13th ACM SIGKDD International Conference on Knowledge Discovery and Data Mining;
2007 Aug 12–15; San Jose, CA, USA; 2007. p. 420–9.
[62] Zhou C, Zhang P, Zang W, Guo L. On the upper bounds of spread for greedy
algorithms in social network influence maximization. IEEE Trans Knowl Data Eng 2015;27(10):2770–83.
[63] Chen W, Wang C, Wang Y. Scalable influence maximization for prevalent viral
marketing in large-scale social networks. In: Proceedings of the 16th ACM SIGKDD International Conference on Knowledge Discovery and Data Mining;
2010 Jul 25–28, Washington, DC, USA; 2010. p. 1029–38.
[64] Goyal A, Lu W, Lakshmanan LVS. SIMPATH: An efficient algorithm for influ-
ence maximization under the linear threshold model. In: Proceedings of the 2011 IEEE 11th International Conference on Data Mining; 2011 Dec 11–14;
Vancouver, BC, Canada; 2012. p. 211–20.
[65] Jung K, Heo W, Chen W. IRIE: a scalable influence maximization algorithm for
independent cascade model and its extensions. Rev Crim 2011;56 (10):1451–5.
[66] Borgs C, Brautbar M, Chayes J, Lucier B. Maximizing social influence in nearly
optimal time. In: Proceedings of the 25th Annual ACM-SIAM Symposium on Discrete Algorithms; 2014 Jan 5–7; Portland, OR, USA; 2014. p. 946–57. [67] Tang Y, Xiao X, Shi Y. Influence maximization: Near-optimal time complexity
meets practical efficiency. In: Proceedings of the 2014 ACM SIGMOD Interna-tional Conference on Management of Data; 2014 June 22–27; Snowbird, UT, USA; 2014. p. 75–86.
[68] Li CT, Lin SD, Shan MK. Influence propagation and maximization for heteroge-
neous social networks. In: Proceedings of the 21st International Conference on World Wide Web; 2012 Apr 16–20; Lyon, France; 2012. p. 559– 60.
[69] Subbian K, Sharma D, Wen Z, Srivastava J. Social capital: The power of influ-
encers in networks. In: Proceedings of the 2013 International Conference on Autonomous Agents and Multi-agent Systems; 2013 May 6–10; Saint Paul, MN, USA; 2013. p. 1243–4.
[70] Li H, Bhowmick SS, Sun A. CINEMA: Conformity-aware greedy algorithm for
influence maximization in online social networks. In: Proceedings of the 16th International Conference on Extending Database Technology; 2013 Mar 18– 22; Genoa, Italy; 2013. p. 323–34.
[71] Lee JR, Chung CW. A query approach for influence maximization on specific
users in social networks. IEEE Trans Knowl Data Eng 2015;27(2):340–53. [72] Deng X, Pan Y, Shen H, Gui J. Credit distribution for influence maximization
in online social networks with node features. J Intell Fuzzy Syst 2016;31
(2):979–90.
[73] Yao Q, Zhou C, Shi R, Wang P, Guo L. Topic-aware social influence minimiza-
tion. In: Proceedings of the 24th International Conference on World Wide Web; 2015 May 18–22; Florence, Italy; 2015. p. 139–40.
[74] Wang B, Chen G, Fu L, Song L, Wang X. DRIMUX: dynamic rumor influence
minimization with user experience in social networks. IEEE Trans Knowl Data Eng 2017;29(10):2168–81.
[75] Groeber P, Lorenz J, Schweitzer F. Dissonance minimization as a microfoun-
dation of social influence in models of opinion formation. J Math Sociol 2014;38(3):147–74.
[76] Chang CW, Yeh MY, Chuang KT. On the guarantee of containment probability
in influence minimization. In: Proceedings of the 2016 IEEE/ACM Internation-al Conference on Advances in Social Networks Analysis and Mining; 2016 Aug 18–21; San Francisco, CA, USA; 2016. p. 231–8.
[77] Faisan MM, Bhavani SD. Maximizing information or influence spread using
flow authority model in social networks. In: Proceedings of the 10th Interna-tional Conference on Distributed Computing and Internet Technology; 2014 Feb 6–9; Bhubaneswar, India; 2014. p. 233–8.
[78] Subbian K, Aggarwal C, Srivastava J. Content-centric flow mining for influence
analysis in social streams. In: Proceedings of the 22nd ACM International Conference on Information & Knowledge Management; 2013 Oct 27–Nov 1;
San Francisco, CA, USA; 2013. p. 841–6.
[79] Kutzkov K, Bifet A, Bonchi F, Gionis A. STRIP: Stream learning of influence
probabilities. In: Proceedings of the 19th ACM SIGKDD International Confer-ence on Knowledge Discovery and Data Mining; 2013 Aug 11–14; Chicago, IL, USA; 2013. p. 275–83.
[80] Teng X, Pei S, Morone F, Makse HA. Collective influence of multiple spreaders
evaluated by tracing real information flow in large-scale social networks. Sci Rep 2016;6(1):36043.
[81] Chintakunta H, Gentimis A. Influence of topology in information flow in social
networks. In: Proceedings of the 2016 Asilomar Conference on Signals, Sys-tems and Computers; 2016 Nov 6–9; Pacific Grove, CA, USA; 2017. p. 67–71. [82] Subbian K, Sharma D, Wen Z, Srivastava J. Finding influencers in networks us-
ing social capital. In: Proceedings of the 2013 IEEE/ACM International Confer-ence on Advances in Social Networks Analysis and Mining; 2013 Aug 25–28;
Niagara Falls, ON, Canada; 2013. p. 592–9.
[83] Liu G, Zhu F, Zheng K, Liu A, Li Z, Zhao L, et al. TOSI: a trust-oriented social
influence evaluation method in contextual social networks. Neurocomputing 2016;210:130–40.
[84] Deng X, Pan Y, Wu Y, Gui J. Credit distribution and influence maximization in
online social networks using node features. In: Proceedings of the 2015 12th International Conference on Fuzzy Systems and Knowledge Discovery; 2015 Aug 15–17; Zhangjiajie, China; 2016. p. 2093–100.
[85] Zhu Y, Wu W, Bi Y, Wu L, Jiang Y, Xu W. Better approximation algorithms for
influence maximization in online social networks. J Comb Optim 2015;30
(1):97–108.
[86] He J, Hu M, Shi M, Liu Y. Research on the measure method of complaint theme
influence on online social network. Expert Syst Appl 2014;41(13):6039–46. [87] Gehrke J, Ginsparg P, Kleinberg J. Overview of the 2003 KDD cup. ACM SIGK-
DD Explor Newslett 2003;5(2):149–51.
[88] Yang J, Leskovec J. Defining and evaluating network communities based on
ground-truth. Knowl Inf Syst 2015;42(1):181–213.
[89] Tang J, Zhang J, Yao L, Li J, Zhang L, Su Z. ArnetMiner: Extraction and mining
of academic social networks. In: Proceedings of the 14th ACM SIGKDD Inter-national Conference on Knowledge Discovery and Data Mining; 2008 Aug 24–27; Las Vegas, NV, USA; 2008. p. 990–8.
2018 工程项目社会评价方法试卷 94分
1.单选题【本题型共15道题】 1.社会评价中的（），也应尽量引用直接或间接的数据，以便对拟建项目社会影响的性质和影响程度给出定性的判断结论。（） A．情景分析 B．定性分析 C．定量分析 D．因果分析 2.社会性别监测与评价中，（）的目的是为项目管理小组提供关于调整项目活动、资源分配及内部管理结构的建议。（） A．项目进度监测 B．社会影响监测 C．项目财务监测 D．社会舆论监测 3.2.13经济合作与发展组织以一个国家或地区社会中位收入或平均收入的（）作为该国家或地区的贫困线。（） A．20% B．30% C．50% D．60% 4.开展公众参与，必然要有时间、资金和技术等方面的投入，并可能带来各种风险。以下不属于公众参与带来的风险的是（）。（） A．可能抬高实际上难以满足的公众对项目的期望值
B．易于引起公众广泛注意，并可能带有感情色彩 C．根据当地人的需求改变项目方案，增加建设成本 D．更容易引起人们对当地不公平现象的特别关注 5.（）有利于为决策者提供项目实施的重要信息，并为项目实施效果评价提供依据。（） A．参与式评价 B．参与式行动 C．参与式监测 D．参与式观察 6.（）是指那些收入在绝对贫困线以上但很接近，并且当受到外部影响时，其收入水平极易下降到贫困线以下的人口。（） A．绝对贫困人口 B．重度贫困人口 C．相对贫困人口 D．临界贫困人口 7.男性和女性所扮演的众多社会角色中，（）是指通过商品生产或家务劳动来获得现金或实物酬劳等的工作。（） A．生产角色 B．繁衍角色 C．服务角色 D．社会文化角色 8.在我国城市地区，（）负责管理社会福利项目，为低收入的城市贫困者提供帮助。（）
公司人才评价体系方案
公司人才评价方案 一、人才评价目的 建立人才评价体系是对公司人才进行客观的评价，通过评价，准确掌握人力资源数量、质量，为公司人才发展、人才使用、人才储备提供可靠依据，实现人力资源统筹配置、人力资本不断增值，更好地为公司的发展提供人力资源支持。 二、人才评价范围 公司范围内转正后满半年工龄的在职管理族员工、专业族员工、技术族员工、营销族员工、操作族员工。 三、人才评价周期 每半年开展一次。每年1月份、7月份各组织一次。 四、人才评价原则 4.1 客观、公正原则。 4.2 定期化、制度化原则。 4.3 可行性、实用性原则。 五、人才评价组织 5.1 资源管理部 5.1.1制订人才评价体系方案； 5.1.2拟订人才评价工作计划、制定评价标准、组织专业测评、审核评价结果； 5.1.3负责人才评价结果的运用（人才任用、人才培养、人才调配）。 5.1.4与部门负责人沟通，保证人才评价工作开展； 5.1.5负责参与人才评价工作人员的培训、工作指导。 5.2用人部门负责人 5.2.1根据评价方案执行本部门人才的评估； 5.2.2提供本部门各岗位的岗位要求，配合建立岗位素质评价模型。 六、岗位评价模型建立 6.1.1 确定评价指标。设计调查问卷,在公司范围内抽取不同层级、不同岗位的员工进行调查，调查员工对公司企业文化的认同、价值观的理解及表现的行为特征，分析调查结果，找出公司人才的核心素质指标。(如：价值观、个性特质)。通过岗位分析，
对完成岗位工作目标所需要的知识、技能进行分析，提炼出各岗位素质指标。可以从以下几方面考虑： 知识。如完成工作目标所需要的专业知识、行业知识、本部门相关知识等； 能力。如为实现工作目标应具有的操作技能、逻辑思维能力、管理能力、沟通能力等。 职业素养。如员工的工作热情、诚信、职业道德等。 6.1.2 确定评价标准。评价标准是对评价指标进行分等级可测量性的描述，体现在行为特征和目标完成结果。分为优秀、良好、一般、较差。 6.1.2.1在公司范围内找出被评价岗位在职员工中优秀员工、一般员工、较差员工进行调查、访谈，在评价指标中具有哪些素质特征和行为特征及工作绩效。通过分析，找出胜任素质指标及行为特征并进行描述，为确定评价的标准提供依据。 6.1.2.2 找出行业标杆企业岗位员工胜任能力模型进行对比，找出差距，按照企业发展的现状和发展方向对评价标准进行调整。 6.1.2.3 成立专家小组,在分析调查数据的基础上,对评价指标、标准进行修正调整。 岗位评价模型的建立要考虑评价的信度和效度，确保评价的准确性、有效性。 七、人才评价方法、程序 7.1人才评价实行自我评价、360°反馈、专项测评机制。 7.1.1 自我评价。 被评价对象按照《岗位评价指标和标准》所列项目自我评价，采用填写《自我鉴定表》的方法进行。 7.1.2 360°反馈。 根据员工所处的职位，选择其直接领导、下属，平级其他部门同僚（员工）对《岗位评价指标和标准》中的项目进行评价。 360°反馈采取问卷法进行评价。 7.1.3 专项测评。 《岗位素质评价指标》中的专业知识、管理能力、个性特征等项目评价难度较大，由资源管理部按照不同职位类别进行专项测评。 专业知识采取试题考试方式进行测评。管理能力、个性特征利用专业的测评工
2018工程项目社会评价方法
2018工程项目社会评价方法 一、单选题【本题型共15道题】 1.社会评价中的（），也应尽量引用直接或间接的数据，以便对拟建项目社会影响的性质和影响程度给出定性的判断结论。（） A．情景分析 B．定性分析 C．定量分析 D．因果分析 用户答案：[B] 得分：2.00 2.（）问题不需要在项目准备阶段的少数民族问题分析时予以重点关注。（） A．少数民族语言 B．调查样本的选取 C．项目监测与评价 D．社会性别分工 用户答案：[D] 得分：2.00 3.社会评价有利于减少项目引发的社会矛盾和纠纷，防止可能产生的 （），促进社会稳定。（） A．风险效应 B．负面效应
C．木桶效应 D．近因效应 用户答案：[B] 得分：2.00 4.（）是指那些收入在绝对贫困线以上但很接近，并且当受到外部影响时，其收入水平极易下降到贫困线以下的人口。（） A．绝对贫困人口 B．重度贫困人口 C．相对贫困人口 D．临界贫困人口 用户答案：[D] 得分：2.00 5.（）是指在项目的设计、决策及实施中具有重要影响的项目利益相关主体。（） A．项目受益人 B．受影响群体 C．关键利益相关者 D．项目目标群体 用户答案：[C] 得分：2.00 6.对于那些涉及到少数民族等问题的工程项目，最好应在（）阶段就启动社会评价工作。（） A．项目准备
B．项目设计 C．项目实施 D．项目运营 用户答案：[A] 得分：2.00 7.（）通过项目所在地区的实际社会状况指标与当地居民对拟建项目的预期进行对比，分析项目可能造成的社会影响及其满足预期的程度。（） A．“有无”对比分析 B．对照组比较 C．现状预期对照比较 D．情景模拟分析 用户答案：[B] 得分：0.00 8.社会性别监测与评价中，（）的目的是为项目方案的调整中，提供需要考虑性别差异的各项建议。（） A．工程进度监测 B．社会影响监测 C．项目财务监测 D．社会舆论监测 用户答案：[B] 得分：2.00 9.我国传统的扶贫开发项目，对受益对象主要采用（）的机制。（）
投资项目的评价方法(一)
投资项目的评价方法（一） 第二节投资项目的评价方法 （一）净现值（NPV）法 特定项目未来现金流入的现值与未来现金流出的现值之间的差额，它是评价项目是否可行的最重要的指标。按照这种方法，所有未来现金流入和流出都要用资本成本折算现值，然后用流入的现值减流出的现值得出净现值。如果净现值为正数，表明投资报酬率大于资本成本，该项目可以增加股东财富，应予采纳。如果净现值为零，表明投资报酬率等于资本成本，不改变股东财富，没有必要采纳。如果净现值为负数，表明投资报酬率小于资本成本，该项目将减损股东财富，应予放弃。 计算净现值的公式： 2.净现值的决策规则：接受“净现值＞0”的项目 3.净现值的原理（超额收益、价值增量） （1）净现值与报酬率的关系 ①净现值＞0，投资报酬率＞资本成本 ②净现值=0，投资报酬率=资本成本 ③净现值＜0，投资报酬率＜资本成本（2）净现值与价值创造——净现值表明投资项目带给投资者的财富增量（按现金流量计量的净收益现值），即对企业来讲： ①净现值＞0，该项目增加股东财富； ②净现值=0，该项目不改变股东财富； ③净现值＜0，该项目减损股东财富。 4.净现值的局限性：绝对值指标，在比较投资额不同的项目时有一定的局限性。 【教材例5-1】设企业的资本成本为10%，有三项投资项目。有关数据如下表。 净现值（A）=（11800×0.9091+13240×0.8264）-20000=1669（元） 净现值（B）=（1200×0.9091+6000×0.8264+6000×0.7513）-9000=1557（元） 净现值（C）=4600×2.487-1200=-560（元）
兼容多个综合评价方案及其分类的数学模型
系统工程学报 JOURNAL OF SYSTEMS ENGINEERING 1999年第14卷第2期 Vol.14 No.2 1999 兼容多个综合评价方案及其分类的数学模型 傅荣林秦寿康陈湛本 摘要本文研究的模型，是与多种综合评价方法的样品排序有最大相关和给定样品排序分类时有最小差异的新评价模型.这些模型已应用于综合评价广州市工业企业50强中. 关键词：评价方案，兼容度，差异度，分类，数学模型 分类号：N94 MATHEMATICAL MODELS OF THE COMPATIBILITY WITH MULTI-VALUATION SCHEMES AND CLASSIFICATIONS Fu Ronglin Qin Shoukang Chen Zhanben (Guangzhou Municipal Institute of Systems Engineering,Guangzhou 510400) Abstract In this paper， we study the new priority methods which have the maximal relatio n and the minimal difference degree with multi-valuation priority methods under g iv en classifications,and the methods are applied in comprehensive evaluation fifty mighty works of Guangzhou City. Key words:valuation schemes,compatibility degree,difference degree,classification，mathema tical model 0 引言 指标体系的综合评价方法是否可靠和准确取决于很多因素，一般来说，不存在普遍适用的综合评价方法.正因为如此，人们业已对系统综合评价方法做了很多研究，提出了许多有效的综合评价方法，如层次分析法、主成分分析法、模糊综合评判法、综合指数法和功效评分法等.这些方法都有各自的优点和特色：层次分析法模型具有层次结构，利于将决策者的经验判断给予量化，对目标结构复杂且缺乏一些数据的情况下更为实用，在社会、经济等领域都有着广泛的应用；主成分分析法模型则是理论上比较成熟，能用少数无关的主成分来代表原来众多相关的指标变量，且可从中提取权向量，因而日益为人们所重视；模糊综合评判法模型能把所有影响对象的独立因素联系到一起，应用等级隶属函数的方法，不仅可给出模糊对象的具体量数据乃至综合评价分，且可判定对象的优劣等级；综合指数法通常能与评价指标的统计口径一致，在社会经济统计中应
咨询师继续教育工程项目社会评价方法试题(卷)答案解析(90分)
一、单选题【本题型共15道题】 1.2.34（）要分析没有拟建项目情况下的当地社会状况，并预测项目建设对当地社会状况的影响，通过对比分析，确定拟建项目的社会影响性质和程度。（） A．“有无”对比分析 B．对照组比较 C．现状预期对照比较 D．情景模拟分析 用户答案：[A] 得分：2.00 2.（）一般以定性描述和分析为主，大致了解拟建项目所在地区社会环境基本状况，粗略预测可能对项目的影响程度。（） A．初步社会评价 B．详细社会评价 C．社会监测评价 D．社会风险评价 用户答案：[A] 得分：2.00 3.参与式的方式强调（）的学习，强调充分了解当地的各种现象从而为发展实践提供可持续的知识基础。（）
A．单向性 B．双向性 C．专业性 D．科学性 用户答案：[B] 得分：2.00 4.（）更偏重于让项目的利益相关者在项目决策和实施等具体活动中发挥作用。（） A．传统行动 B．传统评价 C．参与式行动 D．参与式评价 用户答案：[C] 得分：2.00 5.（）是将项目区与非项目区的社会发展指标进行对比分析，以便揭示拟建项目所带来的实际影响效果的社会评价方法。（） A．“有无”对比分析 B．对照组比较 C．现状预期对照比较
D．情景模拟分析 用户答案：[B] 得分：2.00 6.在我国城市地区，（）负责管理社会福利项目，为低收入的城市贫困者提供帮助。（） A．地方党委 B．民政部门 C．社会保障部门 D．扶贫开发领导小组 用户答案：[B] 得分：2.00 7.2.13经济合作与发展组织以一个国家或地区社会中位收入或平均收入的（）作为该国家或地区的贫困线。（） A．20% B．30% C．50% D．60% 用户答案：[C] 得分：2.00 8.在利益相关者分析中，针对具有较高影响力和较低重要性的利益相关者，应该（）。（）
建设项目经济评价方法
附件二 建设项目经济评价方法 1 总则 1.1 为适应社会主义市场经济发展的需要，规范建设项目经济评价工作，保证经济评价的质量，提高项目决策的科学化水平，引导和促进各类资源的合理有效配置，充分发挥投资效益，制定本建设项目经济评价方法。 1.2 建设项目经济评价方法适用于各类建设项目前期研究工作 (包括规划、机会研究、项目建议书、可行性研究阶段)，项目中间评价和后评价可参照使用。 1.3 建设项目经济评价是项目前期研究工作的重要内容，应根据国民经济与社会发展以及行业、地区发展规划的要求，在项目初步方案的基础上，采用科学、规范的分析方法，对拟建项目的财务可行性和经济合理性进行分析论证，做出全面评价，为项目的科学决策提供经济方面的依据。 1.4建设项目可从不同的角度进行分类。按项目的目标，分为经营性项目和非经营性项目;按项目的产出属性 (产品或服务)，分为公共项目和非公共项目;按项目的投资管理形式，分为政府投资项目和企业投资项目;按项目与企业原有资产的关系，分为新建项目和改扩建项目;按项目的融资主体，分为新设法人项目和既有法人项目。 1.5建设项目经济评价的内容及侧重点，应根据项目性质、项目目标、项目投资者、项目财务主体以及项目对经济与社会的影响程度等具体情况选择确定(见附录A)。 1.6建设项目经济评价的深度，应根据项目决策工作不同阶段的要求确定。建设项目可行性研究阶段的经济评价，应系统分析、计算项目的效益和费用，通过多方案经济比选推荐最佳方案，对项目建设的必要性、财务可行性、经济合理性、投资风险等进行全面的评价。项目规划、机会研究、项目建议书阶段的经济评价可适当简化。 1.7建设项目经济评价必须保证评价的客观性、科学性、公正性，通过 "有无对比"坚持定量分析与定性分析相结合、以定量分析为主和动态分析与静态分析
2016年咨询工程师继续教育工程项目社会评价方法试卷及答案88
2016年咨询工程师继续教育工程项目社会评价方法试卷及答案-88 一、单选题【本题型共15道题】 1.（）主要用于展示当地自然资源、人口、主要经济活动等社会经济信息。（） A．季节历 B．贫困原因分析—因果关系图 C．社会性别劳动分工图 D．社区资源图 用户答案：[D] 得分：2.00 2.在利益相关者的具体构成中，项目的建设单位通常属于（）。（） A．项目受益人 B．项目受害人 C．项目受影响人 D．其他利益相关者 用户答案：[D] 得分：2.00 3.小型项目有可能做到对所有可能受到影响的机构或个人进行调查，但对于大型项目而言，（）可能是获取相关信息的最为有效方法。（） A．情景分析 B．随机抽样 C．蒙特卡洛模拟 D．问卷调查 用户答案：[B] 得分：2.00 4.在（）阶段，要为项目方案设计和实施提供各种社会信息，进行详细社会评价。（） A．项目识别 B．项目准备 C．项目实施 D．项目后评价 用户答案：[B] 得分：2.00 5.（）是将项目区与非项目区的社会发展指标进行对比分析，以便揭示拟建项目所带来的实际影响效果的社会评价方法。（） A．“有无”对比分析
B．对照组比较 C．现状预期对照比较 D．情景模拟分析 用户答案：[B] 得分：2.00 6.（）问题不需要在项目准备阶段的少数民族问题分析时予以重点关注。（） A．少数民族语言 B．调查样本的选取 C．项目监测与评价 D．社会性别分工 用户答案：[D] 得分：2.00 7.2.34（）要分析没有拟建项目情况下的当地社会状况，并预测项目建设对当地社会状况的影响，通过对比分析，确定拟建项目的社会影响性质和程度。（） A．“有无”对比分析 B．对照组比较 C．现状预期对照比较 D．情景模拟分析 用户答案：[A] 得分：2.00 8.参与式分析方式强调分析的过程及结论应该属于被分析的对象，而不是分析者。在进行发展援助时，（）应该成为分析的主体及所有者。（） A．援助者 B．援助的对象 C．利益相关者 D．项目所有者 用户答案：[B] 得分：2.00 9.（）通过项目所在地区的实际社会状况指标与当地居民对拟建项目的预期进行对比，分析项目可能造成的社会影响及其满足预期的程度。（） A．“有无”对比分析 B．对照组比较 C．现状预期对照比较 D．情景模拟分析
基于微博影响力的评价模型(最终版)
基于微博影响力的评价模型 摘要 本文研究的是微博用户和微博的影响力，以及最大好友圈和消息最佳发布问题。 对于问题一，首先，我们查阅相关文献，基于已给数据将用户粉丝数量，用户和粉丝的活跃度，粉丝的专注度作为我们评价微博用户影响力的指标。根据这些指标，我们对题目所给表格中的数据进行统计，将得到的结果作为各个用户在各指标下得到的分数。然后，我们建立了熵权模型，将这些数据进行标准化处理，对各个指标的权重进行了计算，分别得到了在不同人数的微博环境下各指标在评价用户影响力时所占权重。最后，将各指标下分数与权重相乘，再将各个用户的所有指标分数求和得到最终用户影响力评分。将其排序筛选后，得到影响力最大的十个大V编号。 对于问题二，首先，在问题一所建立模型的基础上，我们确立了将微博被转发次数和转发人的影响力作为我们评价微博影响力的指标。然后，我们统计了各条微博在各指标下的数据。其次，我们利用熵权模型将数据标准化处理，并求出了这两个指标在评价微博影响力时所占权重。最后，我们将各微博各指标下分数与权重相乘，并将所有指标分数求和得到最终微博影响力评分。将其排序筛选后，得到影响力最大的十条微博消息。 对于问题三，首先，我们根据题目对于好友圈的定义，将M={(i，j)}定义为关注情况矩阵，即第j用户关注第i用户的情况。进而将多人相互关注的关系用函数关系式表达出来，将所有符合条件的用户编号写入矩阵E中。然后，先通过对符合两两关注情况的用户进行筛选，再从筛选好的用户中逐步添加用户，判断是否为三人相互关注，四人相互关注，直至好友圈内不能再添加人为止。此时的好友圈为最大好友圈。最后，我们通过循环算法实现上述过程，得到了2000人微博环境下10人为最大好友圈，10000人微博环境下13人为最大好友圈，并求出了好友圈中所有用户的编号。 对于问题四，我们将其转化为优化问题，建立了0-1规划模型处理。首先，将用户发布消息与微博用户之间相互关注的情形转化为0-1矩阵（0表示用户未发布消息或表示微博用户之间未相互关注，1则反之），分别表示为x和A。这样若某用户能看到消息，则x与A转置后的矩阵乘积大于等于1。然后，我们将其推广至所有用户，将该问题转化为目标函数为发布消息的人数最少，约束条件为x与A转置后的乘积全部大于等于1的优化问题。最后，我们通过matlab中自带的遗传算法求得最小用户数。但发现效果并非很好。又采用了贪婪算法，求得了2000人微博环境下发布该消息的用户数最少为93人，10000人微博环境下则为249人。 本文建立的模型在已有文献的基础上有一定的改进，求解算法科学，定位速度更快，定位精度更高，有一定的实用和推广价值。 关键词：微博影响力熵权模型遗传算法贪婪算法
2017工程项目社会评价方法试卷
工程项目社会评价方法试卷 一、单选题【本题型共15道题】 1.参与式社会评价强调邀请不同方面的利益相关者参与到社会评价的具体过程中，并特别要求当地人员（）社会评价工作小组。（） A．直接参与 B．间接参与 C．保持回避 D．积极监督 用户答案：[A] 得分：2.00 2.（）主要用于展示当地自然资源、人口、主要经济活动等社会经济信息。（） A．季节历 B．贫困原因分析—因果关系图 C．社会性别劳动分工图 D．社区资源图 用户答案：[D] 得分：2.00
3.当投资项目涉及到少数民族问题时，（）不属于可能导致社会评价需要花费较高费用的因素。（） A．需要考虑语言翻译问题 B．要求具有特殊专长的人才 C．沟通协调相对比较耗时 D．需要更高的补偿标准 用户答案：[D] 得分：2.00 4.社会性别监测与评价中，（）的目的是为项目管理小组提供关于调整项目活动、资源分配及内部管理结构的建议。（） A．项目进度监测 B．社会影响监测 C．项目财务监测 D．社会舆论监测 用户答案：[A] 得分：2.00
5.（）有利于为决策者提供项目实施的重要信息，并为项目实施效果评价提供依据。（） A．参与式评价 B．参与式行动 C．参与式监测 D．参与式观察 用户答案：[C] 得分：2.00 6.（）不是社会评价中项目后评价的主要内容。（） A．社会影响评价 B．项目与社会相互适应性分析 C．社会可持续性分析 D．财务效果评价 用户答案：[D] 得分：2.00 7.社会性别监测与评价中，（）的目的是为项目方案的调整中，提供需要考虑性别差异的各项建议。（）
A．工程进度监测 B．社会影响监测 C．项目财务监测 D．社会舆论监测 用户答案：[A] 得分：0.00 8.（）主要是在性别分析的框架之内确定性别问题，并分析其因果关系的一套操作方法。（） A．意愿调查评估法 B．逻辑框架分析法 C．问题树分析法 D．性别分析方法 用户答案：[D] 得分：2.00 9.在（）阶段，要为项目方案设计和实施提供各种社会信息，进行详细社会评价。（） A．项目识别 B．项目准备
项目评估的方法
根据项目的类型不同，采用的评估方法也不同。 (1)项目评估法和全局评估法。 项目评估法（局部评估法）以具体的技术改造项目为评估对象。赞用、效益的计量范围仅限于项目本身。适用于关系简单，费用、效益容易分离的技术改造项目。例如，投入一笔资金将高耗能设备更换为低能耗设备，只要比较投资和节能导致的费用节约额便能计算出节能的经济效果。 企业评估法（全局评估法）从企业全局出发，通过比较一个企业改造和不改造两个不同方案经济效益变化来评估项目的经济效益。该法既考虑了项目自身的效益，又考虑了给企业其他部分芾来的相关效益。适用于生产系统复杂，效益、费用不好分离的技术改造项目。 (2)总量评估法和增量评估法。 总量评估法的费用、效益测算采用总量数据和指标，确定原有固定资产重估值是估算总投资的难点。该法简单，易被入门接受，侧重经济效果的整体评估，但无法准确回答新增投入资金的经济效果。例如，针对一个小炼钢厂，需要作出是进一步进行技术改造还是关、停、井、转的决策。该项目需要从整体上把握经济效益的变化和能够达到的经济效益指标。此时，应该采用总量法。 增量法采用增量数据和指标并满足可比性原则。这种方法实际上是把“改造”和“不改造”两个方案转化为一个方案进行比较，利用方案之间的差额数据来评价追加投资的经济效果。它虽不涉及原有固定资产重估问题，但却充分考虑了原有固定资产对项目的影响。 增量法又分为前后法和有无法。两者的区别是：前后法使用项目改造后各年的费用和效益减去某一年的费用和效益的增量数据来评估项目改造的经济效益。有无法强调“有项目”和“无项目”两个方案在完全可比的条件下进行全面对比，对两个方案的未来费用、效益均要进行预测并计算改造带来的增量效益。实质上，前后法是有无法的一个特例：即，假定该项目如果不改造，茌未来若干年内经营状况保持不变。这实际上是不可能的，一个企业的经济效益总是在变化的，不是上升，就是下降。因此，一般技术改造项目（包括扩能）+评价都应采用有无法。 (3)费用效益分析法。 费用效益分析法主要是比较为项目所支出的社会费用（即国家和社会为项目所付出的代价）和项目对社会所提供的效益，评估项目建成后将对社会做出的贡献程度。最重要的原则是项目的总收入必须超过总费用，即效益与费用之比必须大于l。 (4)成本效用分析法。 效用包括效能、质量、使用价值、受益等，这些标准常常无法用数量衡评，且不具可比性，因此，评价效用的标准很难用绝对值表示。通常采用移动率、利用率、保养率和可靠程度等相对值来表示。成本效用分析法主要是分析效用的单位成本，即为获得一定的效用而必需耗费的成本，以及节约的成本，即分析净效益。若有功能或效益相同的多项方案，自然应选用单位成本最低者。 成本效用分析有三种情况。
2018年工程项目社会评价方法试卷及答案-98分
1.单选题【本题型共15道题】 1.（）机制要求采用低于市场的工资标准，以便自动地将那些容易找到较高报酬工作的富人排除在外，使参与者主要来自贫困家庭。（） A．区域瞄准 B．家庭瞄准 C．个人瞄准 D．自我瞄准 2.社会监测与评价方案的制定，应重视确定（），分析没有该项目可能会出现的情况。（） A．项目建设情况 B．无项目情况 C．项目技术方案 D．项目社会环境 3.（）更偏重于让项目的利益相关者在项目决策和实施等具体活动中发挥作用。（） A．传统行动 B．传统评价 C．参与式行动 D．参与式评价 4.（）是指那些收入在绝对贫困线以上但很接近，并且当受到外部影响时，其收入水平极易下降到贫困线以下的人口。（） A．绝对贫困人口 B．重度贫困人口 C．相对贫困人口
D．临界贫困人口 5.（）是将项目区与非项目区的社会发展指标进行对比分析，以便揭示拟建项目所带来的实际影响效果的社会评价方法。（） A．“有无”对比分析 B．对照组比较 C．现状预期对照比较 D．情景模拟分析 6.参与式分析方式强调分析的过程及结论应该属于被分析的对象，而不是分析者。在进行发展援助时，（）应该成为分析的主体及所有者。（） A．援助者 B．援助的对象 C．利益相关者 D．项目所有者 7.社会性别监测与评价中，（）的目的是为项目管理小组提供关于调整项目活动、资源分配及内部管理结构的建议。（） A．项目进度监测 B．社会影响监测 C．项目财务监测 D．社会舆论监测 8.（）通过项目所在地区的实际社会状况指标与当地居民对拟建项目的预期进行对比，分析项目可能造成的社会影响及其满足预期的程度。（） A．“有无”对比分析 B．对照组比较
工程项目社会评价方法100分
一、单选题【本题型共15道题】 1.()不是社会评价中项目后评价的主要内容。() A.社会影响评价 B.项目与社会相互适应性分析 C.社会可持续性分析 D.财务效果评价 用户答案：［D］得分:2.00 2.()主要用于展示当地自然资源、人口、主要经济活动等社会经济信息。() A.季节历 B.贫困原因分析一因果关系图 C.社会性别劳动分工图 D.社区资源图 用戸答案:［D］得分:2.00 3?下列不属于项目识别阶段初步社会筛选工作内容的是( )O () A.明确项冃的冃标和主要社会影响 B.明确项目的目标群体与受影响群体 C.评估目标群体对拟建项目的需求 D.评估目标群体对项目的影响力 用户答案：［D］得分：2.00 4.( )是指在项目的设计.决策及实施中具有重要影响的项目利益相关主体匚() A.项目受益人 B.受影响群体 C.关键利益相关者 D.项目目标群体 用户答案:［C］得分：2.00
A. 项目建设情况 B. 无项目情况 C. 项目技术方案 D. 项目社会环境 用户答案：［B ］得分：2.00 6. 社会性别监测与评价中，（）的目的是为项目管理小组 提供关于调整项目活动、资 源分配及内部管理结构的建议。O A. 项目进度监测 B. 社会影响监测 C. 项目财务监测 D. 社会舆论监测 用户答案：［A ］得分：2.00 7. （）通过项目所在地区的实际社会状况指标与当地居民对拟建项目的预期进行对比, 分析项目可能造成的社会影响及其满足预期的程度。（） A. “有无”对比分析 B. 对照组比较 C. 现状预期对照比较 D ?情景模拟分析 用户答案：［C ］得分：2.00 &参与式分析方式强调分析的过程及结论应该属于被分析的对彖，而不是分析者。在进行发展 援助时，（ ）应该成为分析的主体及所有者。（） A. 援助者 B. 援助的对彖 C. 利益相关者 5?社会监测与评价方案的制左，应重视确泄( () 分析没有该项目可能会出现的情况。
《工程项目社会评价方法》继续教育答案
工程项目社会评价方法试卷答案 【试卷总题量: 38，总分: 100.00分】用户得分：71.0分，用时409秒，通过字体：大中小| 打印| 关闭| 一、单选题【本题型共15道题】
1.2.34（）要分析没有拟建项目情况下的当地社会状况，并预测项目建设对当地社会状况的影响，通过对比分析，确定拟建项目的社会影响性质和程度。（） A．“有无”对比分析 B．对照组比较 C．现状预期对照比较 D．情景模拟分析 用户答案：[C] 得分：0.00 2.（）是指在项目的设计、决策及实施中具有重要影响的项目利益相关主体。（）
A．项目受益人 B．受影响群体 C．关键利益相关者 D．项目目标群体 用户答案：[C] 得分：2.00 3.（）更偏重于让项目的利益相关者在项目决策和实施等具体活动中发挥作用。（） A．传统行动 B．传统评价 C．参与式行动 D．参与式评价 用户答案：[C] 得分：2.00
4.社会性别监测与评价中，（）的目的是为项目方案的调整中，提供需要考虑性别差异的各项建议。（） A．工程进度监测 B．社会影响监测 C．项目财务监测 D．社会舆论监测 用户答案：[B] 得分：2.00 5.在我国城市地区，（）负责管理社会福利项目，为低收入的城市贫困者提供帮助。（） A．地方党委 B．民政部门
C．社会保障部门 D．扶贫开发领导小组 用户答案：[B] 得分：2.00 6.社会评价有利于减少项目引发的社会矛盾和纠纷，防止可能产生的（），促进社会稳定。（） A．风险效应 B．负面效应 C．木桶效应 D．近因效应 用户答案：[B] 得分：2.00
教学质量评价模型研究
毕业设计（论文） 教学质量评价模型研究 Research on the model of the Teaching Quality Evaluation 学院: 理学院 学生姓名: 刘君兰 专业:信息与计算科学 班级: 081022 学号: 08102221 指导老师: 孙海 二零一二年六月
摘要 本文研究的是教学质量评价的模型，首先介绍了教学质量评价的发展背景和意义及评价原则，指出了当前教学质量评价中存在的问题，并且提出了改进。教学质量评价分为学生评价、同行和专家评价、领导评价和教师自我评价四个部分，本文分别对这四种评价作出了分析，指出了现有的评价方案中存在的不足之处，并对其进行了相应的改进，使评价方案更加客观、合理、科学。 评价模型主要针对评价指标体系的确定和评价结果的合理处理两个问题展开分析。首先根据查文献得到一个分三个层次的评价指标体系，运用层次分析法和标度法结合得到各评价指标的权重系数，构造合理的教学质量评价体系。为了最大限度地利用评价信息，我们采用有效区间值代替平均值进行教学质量评价。先统计出三级指标的原始评价数据，然后根据区间值模糊法求得各三级指标的有效区间，接着按照分层次模糊法一层一层地计算上一层指标的有效区间，直到得到最高层次的有效区间，根据四种评价主体得到的有效区间得到综合有效区间，求出该有效区间在各等级的隶属度，根据最大区间数判断最后综合评价的有效区间的隶属等级。 按照改进后的要求，在Matlab GUI环境下进行编程得到一个可视界面，只需在得到的界面中分别输入各评价主体对三级指标的评价分数，就可以得到综合评价的有效区间，判断出该有效区间的隶属等级。 关键词:教学质量评价；层次分析法；标度法；有效区间；GUI可视界面
全国建模比赛一等奖2010年上海世博会影响力的定量评估
2010年上海世博会影响力的定量评估 摘要 本文从四个角度对2010年上海世博会的影响力进行了评估。 第一，从纵向考虑，通过最近四届世博会与上海世博会在参观国家和组织数目、参观人数、场馆数目、持续时间、活动常数等指标上的比较，定义“直接影响力”的概念，采用因子分析法，得出这四届世博会直接影响力的排名： 上海世博会>爱知世博会>汉诺威世博会>萨拉戈萨世博会第二，选取上海世博会对上海市旅游业的短期影响这一侧面，用世博举办前的指标数据进行自回归将世博对上海旅游业的影响从实际数据中剥离，建立旅游本底趋势线模型，用本底值与实际值的差值作为世博的影响值，估算出世博会对上海旅游业从2003年到2012年的时间区间内的贡献率和影响率。 第三，通过对比其它大型国际盛会对其本地旅游业贡献效率，分析上海世博会对旅游业的影响力。建立数据包络分析评估模型，选取投入、主场馆占地面积、入境旅游人数增长百分比和旅游收入增长百分比四个指标得出其技术效率值为0.2911和规模效率值0.2919。 第四，由于世博会的长期效益受时间影响，故建立影响力关于时间的影响力降温模型，并通过对比其它顶级盛事，用积分模型反映出上海世博会未来一段时间内的综合影响力。 关键词：世博会，影响力，本底趋势线模型，数据包络分析
1．问题重述 2010年上海世博会是首次在中国举办的世界博览会。从1851年伦敦的“万国工业博览会”开始，世博会正日益成为各国人民交流历史文化、展示科技成果、体现合作精神、展望未来发展等的重要舞台。请选择感兴趣的某个侧面，建立数学模型，利用互联网数据，定量评估2010年上海世博会的影响力。 2．模型假设和符号说明 2.1模型假设 1）假设所查数据真实可靠； 2）假设忽略短期内各国货币的购买力随时间的变化； 3）假设预计有7000万左右游客参观上海世博会这一数据准确； 4）忽略各国之间消费水平的差异 2.2符号说明 t：年份； M：旅游人数； S：旅游外汇收入； G：国民生产总值； B：旅游收入占国民生产总值的比重； ()t L：影响力降温特性曲线； ()t f：温度下降函数； v：降温速度； T：最小温度； m P：上海世博会在未来一段时间内的综合影响力 s θ：CCR模型效率值； λ：CCR模型中各决策单元权重; i：输入指标； r：输出指标； j：决策单元； - S：输入指标松弛变量； + S：输出指标松弛变量； 3．问题分析 本题要求对2010上海世博会的影响力作出定量评估，要求看似简单，但较为开放，发散性比较大，容易展开但也难以下手。并且影响力本身是一个抽象概念，要对其进行定量评估，必定要选取能够体现其影响力的某个或几个方面，查询相应指标，才能对其做出定量客观的评价。 世博会和奥运会、世界杯一起并称全球三大顶级盛事，其影响力是不言而喻的，因此选取的比较对象必须具备可比性。首先可以想到的是本届世博会同往届世博会的对比，考虑到时间跨度越大，经济、通信、交通等条件的差异就越大，所以尽量选择近几年举办的世博会和上海世博会进行比较，可以从参展的国家和组织、游客人数、场馆规模等指标出发进行对比，评价上海世博会在世界范围内的一个影响力。
(完整版)建设项目经济评价方法与参数(第三版)
建设项目经济评价方法与参数（第三版）（以下简称方法与参数三）主要由建设项目经济评价方法和建设项目经济评价参数两部分组成。其中建设项目经济评价参数主要由指标的计算方法和各指标的标准参考值组成。建设项目经济评价方法包括总则、财务效益与费用估算、资金来源与融资方案、财务分析、经济费用效益分析、费用效果分析、不确定性分析与风险分析、区域经济与宏观经济影响分析、方案经济必选、改扩建项目与并购项目经济评价特点、部分行业项目经济评价的特点。综合起来看，根据方法与参数三中的有关规定，需要对于“工程造价计价与控制”教材的内容做以下调整： 一、流动资金的估算 流动资金估算方法可采用扩大指标估算法或分项详细估算法。分项详细估算法的具体计算公式为： 1. 周转次数的计算： 各类流动资产和流动负债的最低周转天数参照同类企业的平均周转天数并结合项目特点确定，或按部门（行业）规定，在确定最低周转天数时应考虑储存天数、在途天数，并考虑适当的保险系数。 2. 流动资产的估算。 （1）存货的估算。存货是指企业在日常生产经营过程中持有以备出售，或者仍然处在生产过程，或者在生产或提供劳务过程中将消耗的材料或物料等，包括各类材料、商品、在产品、半成品和产成品等。为简化计算，项目评价中仅考虑外购原材料、燃料、其他材料、在产品和产成品，并分项进行计算。计算公式为： （2）应收账款估算。应收账款是指企业对外销售商品、提供劳务尚未收回的资金，计算公式为： （3）预付账款估算。预付账款是指企业为购买各类材料、半成品或服务所预先支付的款项，计算公式为：
（4）现金需要量估算。项目流动资金中的现金是指为维持正常生产运营必须预留的货币资金，计算公式为： 现金=（年工资及福利费+年其他费用）/现金周转次数 年其他费用=制造费用+管理费用+营业费用—（以上三项费用中 所含的工资及福利费、折旧费、摊销费、修理费） 3. 流动负债估算。流动负债是指将在一年（含一年）或者超过1年的一个营业周期内偿还得债务，包括短期借款、应付票据、应付账款、预收账款、应付工资、应付福利费、应付股利、应交税金、其他暂收应付款项、预提费用和一年内到期的长期借款等。在项目评价中，流动负债的估算可以只考虑应付账款和预收账款两项。计算公式为： 应付账款=外购原材料、燃料动力及其他材料年费用/应付账款周转次数 预收账款=预收的营业收入年金额/预收账款周转次数 二、财务基础数据测算表 1. 进行财务效益和费用估算，需要编制下列财务分析辅助报表： （1）建设投资估算表； （2）建设期利息估算表； （3）流动资金估算表； （4）项目总投资使用计划与资金筹措表； （5）营业收入、营业税金及附加和增值税估算表； （6）总成本费用估算表。若用生产要素法编制总成本费用估算表，还应编制下列基础报表： 1）外购原材料费估算表； 2）外购燃料和动力费估算表； 3）固定资产折旧费估算表； 4）无形资产和其他资产摊销估算表； 5）工资及福利费估算表。 2. 财务基础数据测算表之间的相互关系。各财务基础数据测算表之间的关系可如图1所示： 三、生产成本费用估算 1. 总成本费用估算。总成本费用系指在运营期内为生产产品提供服务所发生的全部费用。等于经营成本与折旧费、摊销费和财务费用之和。总成本费用可按下列方法估算： （1）生产成本加期间费用估算法：
数学建模常见评价模型简介 (2)
常见评价模型简介 评价类数学模型是全国数学建模竞赛中经常出现的一类模型，如2005年全国赛A题长江水质的评价问题，2008年B题高校学费标准评价体系问题等。主要介绍三种比较常用的评价模型：层次分析模型，模糊综合评价模型，灰色关联分析模型，以期帮助大家了解不同背景下不同评价方法的应用。 层次分析模型 层次分析法(AHP)是根据问题的性质和要求，将所包含的因素进行分类，一般按目标层、准则层和子准则层排列，构成一个层次结构，对同层次内诸因素采用两两比较的方法确定出相对于上一层目标的权重，这样层层分析下去，直到最后一层，给出所有因素相对于总目标而言，按重要性程度的一个排序。其主要特征是，它合理地将定性与定量决策结合起来，按照思维、心理的规律把决策过程层次化、数量化。 运用层次分析法进行决策，可以分为以下四个步骤： 步骤1 建立层次分析结构模型 深入分析实际问题，将有关因素自上而下分层(目标—准则或指标—方案或对象)，上层受下层影响，而层内各因素基本上相对独立。 步骤2构造成对比较阵 对于同一层次的各元素关于上一层次中某一准则的重要性进行两两比较，借助1~9尺度，构造比较矩阵； 步骤3计算权向量并作一致性检验 由判断矩阵计算被比较元素对于该准则的相对权重，并进行一致性检验，若通过，则最大特征根对应的特征向量做为权向量。
步骤4计算组合权向量(作组合一致性检验) 组合权向量可作为决策的定量依据 通过一个具体的例子介绍层次分析模型的应用。 例(选择旅游地决策问题)如何在桂林、黄山、北戴河3个目的地中按照景色、费用、居住条件、饮食、旅途条件等因素进行选择。 步骤1 建立系统的递阶层次结构 将决策问题分为3个层次：目标层O，准则层C，方案层P；每层有若干元素，各层元素间的关系用相连的直线表示。 图1 选择旅游地的层次结构 步骤2构造比较矩阵 元素之间两两对比，对比采用美国运筹学家A.L.Saaty教授提出的1~9比率标度法(表1)对不同指标进行两两比较，构造判断矩阵。
社会影响力分析——模型、方法和评价
Research Cybersecurity—Review Social In?uence Analysis:Models,Methods,and Evaluation Kan Li ?,Lin Zhang,Heyan Huang School of Computer Science and Technology,Beijing Institute of Technology,Beijing 100081,China a r t i c l e i n f o Article history: Received 10December 2017Revised 5January 2018Accepted 8January 2018 Available online 16February 2018Keywords: Social in?uence analysis Online social networks Social in?uence analysis models In?uence evaluation a b s t r a c t Social in?uence analysis (SIA)is a vast research ?eld that has attracted research interest in many areas.In this paper,we present a survey of representative and state-of-the-art work in models,methods,and eval-uation aspects related to SIA.We divide SIA models into two types:microscopic and macroscopic models.Microscopic models consider human interactions and the structure of the in?uence process,whereas macroscopic models consider the same transmission probability and identical in?uential power for all users.We analyze social in?uence methods including in?uence maximization,in?uence minimization,?ow of in?uence,and individual in?uence.In social in?uence evaluation,in?uence evaluation metrics are introduced and social in?uence evaluation models are then analyzed.The objectives of this paper are to provide a comprehensive analysis,aid in understanding social behaviors,provide a theoretical basis for in?uencing public opinion,and unveil future research directions and potential applications. ó 2018 THE AUTHORS. Published by Elsevier LTD on behalf of Chinese Academy of Engineering and Higher Education Press Limited Company. This is an open access article under the CC BY-NC-ND license 1.Introduction Online social networks such as Weibo,Twitter,and Facebook provide valuable platforms for information diffusion among their users.During this process,social in?uence occurs when a person’s opinions,emotions,or behaviors are affected by other people [1].Thus,changes occur in an individual’s attitudes,thoughts,feelings,or behaviors as a result of interaction with other people or groups.Social in?uence analysis (SIA)is becoming an impor-tant research ?eld in social networks.SIA mainly studies how to model the in?uence diffusion process in networks,and how to propose an ef?cient method to identify a group of target nodes in a network [2].Studied questions include:Who in?uences whom;who is in?uenced;who are the most in?uential users,and so forth.SIA has important social signi?cance and has been applied in many ?elds.Viral marketing [3–10],online recommen-dation [11],healthcare communities [12–14],expert ?nding [15–17],rumor spreading [18],and other applications all depend on the social in?uence effect [19–21].Analyzing social in?uence can help us to understand peoples’social behaviors,provide the-oretical support for making public decisions and in?uencing pub-lic opinion,and promote exchanges and dissemination of various activities [22]. This paper provides a comprehensive view of SIA from the aspects of models,methods,and evaluation.To this end,we iden-tify the strengths and weaknesses of existing models and methods,as well as those of the evaluation of social in?uence.First,we review existing social in?uence models.Next,we summarize social in?uence methods.Finally,we analyze the evaluation of social in?uence. The rest of this paper is organized as follows.In Section 2,we discuss SIA models.In Section 3,we analyze SIA methods,includ-ing in?uence maximization,in?uence minimization,?ow of in?u-ence,and individual in?uence.We then detail social in?uence evaluation in Section 4.Finally,we summarize the reviewed mod-els and methods of social in?uence,and discuss open questions.2.Social in?uence analysis models SIA models have been widely studied in the literature.We clas-sify these models into two categories:microscopic and macro-scopic models.2.1.Microscopic models Microscopic models focus on the role of human interactions,and examine the structure of the in?uence process.The two fre-quently used in?uence analysis models in this category are the independent cascade (IC)[23–25]and linear threshold (LT) ?Corresponding author. E-mail address:likan@https://www.360docs.net/doc/664755022.html, (K.Li). Engineering 4(2018) 40–46 Contents lists available at ScienceDirect Engineering