Research On Web Page Personalized Recommendation Technology

With the fast development and extensive application of Internet technology,the information on the website has been growing exponentially.The website users are satisfied because it meets their informational needs.However,they are also suffering from such huge amount of information,as well as those problems brought by the World Wide Web due to its distributive,dynamic,massive,heterogeneous,complex and open properties.Thus,a novel technique is badly in need,which could automatically dig out desirable information from the huge pool of resources as quick as a flash,withdrawing it and at the same time filtering out other information unwanted.Fortunately,the emergence of personalized recommendation technology relieves us of infinite data and the commercialized world by saving us plentiful time and energy on searching information.In addition,this new technology also successfully changes the service of website from webpage-centered mode to user-oriented one.It supplies users with personalized services and forges ahead toward the realization of supreme level of the whole internet services.In this case,it is of vital significance to find ways to improve existing technology in order to uplift the quality of personalized recommender system.Association-rules-based personalized recommender is one of the most popular forms of recommender technology.In this paper,after detailed analysis of this technology,unreasonable aspects of its general model were found and further optimization work was done successfully.Firstly,the backgrounds and status of this research were presented,followed by general review of web mining and personalized recommender technology.We focused mainly on the association-rule-mining and Apriori algorithm.After in-depth study of general model of association-rule-based personalized recommender system,we found the existing problems in it and proposed several methods of improvement.Finally, given the theoretical analysis and experimental validation,the revised recommender model was proved to have much better quality.