Service-oriented Consumer Behavior Analysis And Recommend Research Model

With the rapid popularization of Internet application in commercial field, the consumer’sdemand becomes the pilot which drives the development of Internet. Meanwhile, the socialnetwork represented by e-commerce varies constantly. In the e-commerce which is highlyintegrated of computer system and social interaction system, the consumers become more andmore important; the consumers’ behavior characteristics and their potential needs graduallybecome the important power to impel the development of technology. Therefore, to researchconsumers decision behavior, to analyze consumers information characteristics, as well as todiscover the underlying rules in consumer behavior and the relational degree between consumers,could guide the computer network develop better and make e-commerce develop towards thedirection of more intelligent and more human, thus mix together with the real social system moreharmoniously.The research on online consumer behavior closely links with many subjects, such as, socialpsychology, computer science, economics, marketing, anthropology, physics and everything todo with complex network. The ultimate goal is to guide the computer network technology,especially e-commerce, providing a better service to the development of human society. A largenumber of enterprises’ collaborative filtering recommendation systems solve such problems asinformation overload, through establishing consumer prediction model and providingpersonalized service. Based on the above requirements, this paper goes further into the researchprogress of service theory and the service quality management theory, analyzingcomprehensively the research origin, research status and research method of service science,excavating profoundly the existing problems of the study on service science. With the servicequality evaluation and satisfaction theory in modern marketing, this research builds thetheoretical research framework serving the dominant logic, providing practical guidance forenterprises to improve the service innovation ability and theoretical direction for systemplanning. The research content includes the following aspects.1. Presenting the research model on consumers buying decision behaviorBased on the initial trust and TAM theory, this paper gives the empirical research on theconsumers first time buying behavior based on the online electronic supplier through interviewsand questionnaire survey. In order to enhance consumer’s intention of the first time buying of theelectronic supplier, the research adds a new variable, namely the perceived service quality for theelectronic suppliers which influence directly the first time purchase intention. According to theresults of the analysis, the research puts forward a vertical theory system.In this theory system, the establishment of initial trust and perceived service quality theoryare critical for the new online consumer’s intention of the first time purchase. According to theresult of the statistical analysis, this research finds the effect of various kinds of influence factorson consumers’ intention behavior, reminding the enterprises to pay attention to the establishmentof initial trust in projecting the websites elements of online shopping system, while consumersperceived service quality cannot be neglected. 2. Analyzing the long tail theory in electronic commerce and applying it to the movierecommendationIn the long tail theory, with the development of information technology, the trend ofeconomic development is from a mixture of pure economics to byte and atoms, then to theeconomy of pure bytes. It could recommend the really satisfying products and services to theusers through the analysis of the potential information in big data.There is no recommendation system which could be applied to all products. Only the userswho are deeply involved in the personalized recommendation, can the system speculates theirpreferences in a certain accuracy and diversity according to the fragments of information, andrecommends to the user according to this. The personalized recommendation technology can notonly save users browse search time, but also can find some potential information in the corner ofthe network, digging out the profit from the long tail theory in the information ocean. On thebasis of deep summary of related issues of the long tail theory, this paper analyzes the long tailtheory application in movie recommendations.3. Proposing and verifying the collaborative filtering recommendation model based ondiversityThe accuracy of the traditional online recommendation system, much depends on thecollaborative filtering recommendation algorithm, however, recommend system aims to attractthe interest of consumers and turn visitors into buyers, rather than accurately predict their score.Online recommendation system is the service version of social filtering process. Mostprevious studies emphasize the accuracy of the collaborative filtering algorithm. However, theeffective recommendation system must be credible. It requires that the system logic betransparency and the system be able to provide consumers a new, inexperienced project. Basedon the above, this paper proposes to research the quality evaluation of recommendation systemfrom the angle of user’s experience, adding a freshness parameters of Top-N recommendcollaborative filtering similarity calculation method, and comparing with the classicalrecommended algorithm. The result of this experiment has a certain degree of accuracy and highdiversity, which provides basis for establishing the e-commercial recommendation system.4. Putting forward the emotion text recognition algorithm based on maximum entropytheoryThe large number of consumers text reviews information has great influence on otherconsumers perception of enterprise’s public praise and reputation, hiding a lot of consumer’spreferences and behavior characteristics. Using the text reviews information analysis technology,the enterprises could know the perceived service quality and preferences of their reputation,goods and services from consumers.This paper put forward an online review emotional collocation recognition algorithm basedon maximum entropy model to recognize the emotional tendency of online reviews. Thisresearch designs maximum entropy model based on semantic features of sentimental words,regarding the synonyms in emotion text review corpus as a semantic feature class according tothe synonym lists, choosing the emotion text review corpus which contains a certain semantic feature to construct atomic feature template and composite feature template based on maximumentropy model, thus for automatic identification. The judgment on the collocation relationshipbetween evaluation object and evaluative words in emotion text review corpus is a dichotomy ofcollocation and non-collocation. This paper dichotomizes the evaluation of collocation based onmaximum entropy model and expands the maximum entropy feature template by establishingpolarity glossary. The experiment proves that the model can improve the accuracy ofclassification.