Based On Web Data Mining Of Agricultural Information Personalized Recommendation Key Technology Research

With the rapid development of Internet, the agricultural information websites is rapid development.However, due to the technical level of agricultural information website users and agricultural information website building issues, the user is difficult in a short period of time to find what you want agricultural information.How to provide users with personalized service has become an important issue in the field need to be resolved.On the basis of papers in a comprehensive analysis of Web mining, agricultural information websites, personalized recommendation system research and development status and problems,combined with the characteristics of the agricultural information websites,proposed a solutions for the based on web data mining of agricultural information personalized recommendation system,depth study of three main problem were Session identification,user classification, personalized recommendation algorithm.the specific work as follows:1. This paper presents an introduction of dynamic degree of interest in the current page time threshold current user session identification. For different users access to different pages, the introduction of the concept of the factor of interest, and to dynamically adjust the time of the session identification threshold value the more accurately identifying the session.2. Base on the user tree browser mode,, this paper put forward a user classification methods based on plant morphology.Morphological thinking mainly on the user’s browser the path tree shape characteristics, different shape features for users to browse the path tree, divided users into four categories were shallow user, knowledge building user, hedonic browsing user,normal user and builded the decision classification tree.3. This paper put forward a recommendation algorithm based on the network structure. The algorithm with the user and resource network structure and user browsing path tree concept combining the morphological characteristics of the definition of the various types of users, so that they use in the process of allocating resources triples data to classify the resources allocated resources recommended consider the type of user, the process to be given a different weight, preferably recommended.4. This paper put forward a agricultural information personalized recommendation system based on web data mining solutions. Experimental comparison proposed algorithm and traditional network-based recommendation algorithm. Prove that the proposed algorithm has certain advantages for different types of users recommended...