Recommendation System Based On User Interest Model Of The Search Engine Results

With the emergence of Web Technology, internet has become main method for people to obtain information. However, facing the overwhelming information provided by internet, people are often in difficult situation of’information overload’. The search engine application may help user solving this problem to some extent, but the current search engine system cannot provide personalized recommendation for users, and are unable to satisfy different users’demand. Proposed a search engine result recommendation system based on web user interest model, aiming to help users to find interested information more quickly and accurately.Firstly, research and study user interest model related technologies, including the obtaining of data, the representation method of user interest, user model building technology and updating technology etc. In the meantime, did some research on traditional search engine system as well.Base on the research and study above, proposed a user interest model, which is single user oriented, and can be built automatically. The user model is using IE browsing history as the data source. The system will pre-process the data, segment the words... build the vector for each web page, finally, form the user interest model by aggregating all the vectors.At last, created a personalized search engine system. The system is based on the user interest model. It will calculate the similarity between each result page and the user interest model, then present those pages having higher similarity with the user model to users. Have simulated each steps of the system and have given the result. The result proves that:the personalized system based on the user interest model can meet users’demand very well, and can greatly improve the efficiency of obtaining the information.There are many places that people can use user interest model for information referral. For example:1) Advertisement delivery, to determine how it’s matching with the user’s interest; 2) Personal information agent. Automatically determine user’s interesting level on certain page, based on user interest model. Researching of the techniques of user interest model has very positive significance on improving the efficiency of internet users.