Research On The Characteristics Of Online Commodity Transaction Behavior Based On Complex Network

Along with the booming of internet technology, the on-line goods trading which is one kind of e-business has developed very rapidly in our country; however, some problems are exposed as well, such as the severe lack of e-business credit, merchants’nonobservance of promises, network fraud, etc. which happen frequently. These problems have severely influenced the development of our country’s e-business. So far, the credit feedback mechanism has become the important assurance for the healthy development of e-business network, and the credit of both parties to on-line trading becomes a key factor that restrict the on-line trading, therefore, it is of important theoretical value and practical meaning to study the problems with goods trading on e-business platform.Actually, the on-line trading process and establishment of both sides’ reputation are influenced by many factors including but not limited to the complexity of the on-line trading platform structure, also the customer consumption pattern and goods trading evaluation mechanism influence the approach of on-line trading to a large extent. In view of the platform structure and characteristics on which the e-business mainly relies, it has distinct social network features, and has the characteristics of complex network structure; on such basis, it is inevitable to study the on-line trading platform mechanism. The small world phenomenon in complex networks is a significant finding in the field of sociology, and it indicates that people of any two social relation networks can be connected via only a few links; after the small world phenomenon being put forward, many network models based on its characteristics are set up, such as the random graph model, Albert-Barabasi (BA) model, Watts-Strogatz(WS) model, etc., providing the tool support to the study of network characteristics in subjects of sociology, physics, economics, etc.Therefore, taking the study of time-varying dynamic model and extended small world model of complex networks as the breakthrough point, and based on the clustering feature of complex networks, this paper carries out the critical-path analysis on the development tendency of network community structure and the transmission mode of customer on-line comments through the algorithm of complex network model, expands and deepens the customer trust transmission mechanism of e-business, as well as studies the main problems in current e-business on-line trading.Firstly, on basis of summarizing the existing documents on complex networks, the paper analyzes the general characteristics of complex network theoretically. Via detailed analysis of complex network, measure and construct the complex networks model in match with the characteristics of e-business network with time variation as the core element, and finally, on basis of deeply analyzing the time varying complex network model, connect with the extended small world model and analyze the characteristics and advantages of such model to form the time-element based complex network theoretic foundation.Secondly, based on the study on the structure of on-line trading community of complex network, this paper studies the community attribute of customer purchasing behavior in e-business; different customers gradually form the stable community structure through on-line trading and develop certain consumption stickness. It also takes advantage of the aggregation characteristic of complex network to study the group behavior characteristic of customers in e-business. Through theoretical review of clustering algorithms of complex network and the comparison between different algorithms, the paper sets up and analyzes the clustering method of on-line trading platform community structure. On basis of the constructed complex network model, integrate such critical variants in e-business trading behavior as credit, trust, customer consensus transmission, etc. into it to set up the e-business complex network model, then dig the community structure of network trading platform by optimizing the existing algorithms, as well as perform experiment by using the computer software so as to verify the reliability of algorithm model with real data.Next, the paper studies the on-line trading information and evaluation opinion transmission. Through theory modeling and simulation analysis, the paper analyzes the complex networks formed by information transmission during customer consuming, and further discusses that the opinions and viewpoints of network community is characterized of spontaneous organization, which is the social foundation necessary for forming network opinion leader; meanwhile, through the identification method of network opinion leader, the paper discusses about the discovery of opinion leader of network community via network structure.Finally, the paper studies the mechanism of customer trust transmission in e-business. It analyzes the generation and transmission process of trust and the importance of establishing enterprise customer reputation relation in e-business; discusses the crucial influence of personalized recommendation and customer interaction in establishing the trust transmission mechanism. On such basis, it sets up a dynamic trust model based on the recommendation, and through analyzing the recommendation process in human society, studies the variant measurement method and specific algorithm, then uses the computer simulation software to perform simulation experiment on the constructed model; the experiment results again have verified the correctness of theoretical analysis and model construction in this paper from the aspect of actual data, offer a connection-based recommendation trust searching mechanism, set up an effective evaluation trust network model, and solve the problem of trust transmission being difficult to be quantized in e-business trading.The paper has made theoretical and practical contributions in the following aspects:First, on basis of setting up the personalized recommendation, make recommendation according to the credit principle, thus study the customer trust transmission mechanism in e-business. It has solved the following"difficult points:the true structure of complex networks cannot be presented fully by simply distributing weight of model with nodes, because there are such problems as asymmetry of purchase amount, individuality of evaluation, the size of purchase amount being differentiated due to influences of time and other relevant elements, degree of good and bad evaluations, etc. The paper carries out details construction without compromising the accuracy much, measures the recommendation trust through the method of aggregating recommendation nodes to give the recommendation values. Through analyzing the recommendation process of human society, the papers provides a connection-based recommendation trust searching mechanism, determines the trust degree of the recommendation value given by each recommendation node, and has obtained good effect.Second, the paper analyzed the transmission mode of on-line trading information and evaluation viewpoints in e-business. It constructs the matrix model of virtual community in correspondence with cell space; by effective implementation of inclination transformation principle, the problem of network opinion spillover is properly solved; that is to say, in order to avoid errors, the paper has calculated and arranged the whole evaluation information of on-line consumers, and divided into the corresponding specific values and degrees; in aspect of inclination transformation, the paper adopts the method of upper, middle and lower value locking and successfully solves this impediment by using formula and spillover function. On such basis, the paper describes such network consumption experience as customers’opinion varying as the time varies and the establishment of evaluation, and establishes the model of dynamic evolution process of opinion transmission and evaluation generation. This model is accurate and effective in reflecting the opinion transmission, consumption experience and evaluation establishment process in on-line trading networks and can truly reflect the network trading process in reality.Third, dig the community structure of network trading platform through optimizing the existing algorithms, and after defining the complex contents and problems of constructing the on-line trading platform, the study also puts forward corresponding model solution program against this. First of all, adopt the priority growth hypothesis specific to the topology structure, lay down the priority growth program according to the known data, thus simplify the topology structure to seek for the key nodes; then play the important role of transmission immune mechanism for information selection and spread, and analyze the overall tendency of information selection through the positive and negative messages and evaluations; finally, use the computer software to simulate the technology to do the experiment, and verify the reliability of algorithm model with real data, and deeply discuss the influence and effect of information communication and relation of on-line trading customers.