The Research Of UGIM Construction Based On Consumer Behavior

With the rapid development of computer technology and the popularity of the Internet, the information age in which humans are rapid changes, which along with the behavior of human beings is more complicated and information service for the higher demands and expectations. The behavior of human beings more irregular data, but hope to get more intimate information services, but from a technical and commercial point of view, this is a contradiction. Therefore, a thorough understanding of user behavior, to determine the degree of interest and points of interest of users, to build accurate global user interest model(UGIM) is the most effective way to resolve this contradiction.So, whether in academia or commercial use, UGIM research has always been a popular and practical field; only to build a scientific and rational UGIM, on this basis in order to provide users with one hand, targeted personalized service, while providing for the operation of the business decision-making more scientific basis, to provide more resources to more valuable customers, so as to solve the contradiction individual needs and complexity of data between high cost, to information providers and Users win-win.This article is based on the above background, graduate education innovation fund projects subject Wuhan University engineering support, interpretation and practical application of the theory combined with machine learning and classification model for research paradigm, a study of the following:1. The summary provides an overview of the theory of user behavior in the network environment. Summary of the traditional definition and classification of user behavior, summarizes the collection and analysis of user behavior. Pointed out the lack of collection and collation of traditional data, and highlights the latest, based on the data network open platform to collect and use.2. "User Global interest model(UGIM)" theory systematic description and analysis, summarized its content, classification, definition, represents the results of relevant research methods, and highlights based clustering algorithm k-nn construction method and ontology-based representation, laid the theoretical basis for the prototype system in this paper.3. Examples of user behavior demonstrates collection and collation of data, based on the k-nn clustering idea and ontological representation method UGIM build process, combined with the actual business thinking, respectively, from the user’s potential consumer interest degree, the user and the user’s point of interest ability and willingness to build three three UGIM, UGIM each category are to be expressed by means of ontology, the calculation process follows the k-nn clustering thought. Experimental results show that, UGIM recall rate and accuracy have reached the desired results, and this thought process is more clear, user behavior indicators considered more fully and effectively to the practical application of research as the guiding ideology, So the research of UGIM is also necessary.4. UGIM use research, mainly on UGIM used in e-commerce system has been studied, on the one hand, through its use of the system can more accurately predict the content and extent of the interest to the user, and can be intelligent recommendation; another also from the point of view of the information providing effective resource available to the more valuable targeted customers.