The Research And Implement On The Personalized Meta Search Engine

How to find the information that the user requires in the voluminous already become the hot topic in information retrieval service domain. Future search service should be able to provide the personalization information service to the user. The personalized modeling which is the core technology of the personalized service's also gradually was taken seriously. The personalized model quality immediate influence to personalized service quality. When the system had understood the user's interest,the user's visit pattern well, it can afford the ideal personalized service for the user. This paper has conducted the research regarding the personalization modeling technology, designed and has realized PMSE (the personalized Meta search engine) system.System aim: Through mining user's interest, this system establishes the reasonable model to express the user interest and optimizes it by unceasingly the renewal and the maintenance. This model can reflect the user's interest accurately. Based on this modal this system can provide the personalized service for the user. Research in detail on PMSE:(1) Referring to the ODP classified level, the system constructs user's interest tree to dynamic storage user's interest and describes the user's interest with the short-term interest and the long-term interest.(2) Has studied the user's interest mining algorithm, we combine the explicit and the implicit mining technology in order to the model can reflect the user's interest well.(3)We withdraw the user's interest subject characteristic from user's visited pages in the implicit mining and proposed an algorithm that extract the feature of homepages and moreover also proposed page classification algorithm based on the personalized model.(4)Study the intention of users'queries, to narrow the scope of the user's finding and improve the accuracy of the query. The paper puts forward a user's query to the user interest mapping algorithm. Users interested in it through a combination of models to speculate user's query intent, analysis of interest to users to search categories, and expand the user query.(5) A Meta search engine on the scheduling strategy. Through comparing the existing research on this topic, we proposed a sub-engine classification based on sampling interested members if the characteristics of the sub-engine and an algorithm based the user's interest.In the PMSE system, the user's interest model is a very important module. Only had the accurate modal, the PMSE provides the information retrieval well for user. And how to construct the user's interest is very important in this model. With the precise, the PMSE can work well. Users interested in the model of the process of mining algorithms have a large proportion of high-quality user only has interest in mining algorithms, in order to build good user model can be achieved personalized service and better meet the needs of users. This paper presents the multi-level model of user's interest mining algorithms, based on the ODP; it has high accuracy and can afford user expectations.