Personalized Search Engine Based On Ontology Research

With the continuous development of the search engine, personalized search has gradually become an efficient search tools for the users who have different knowledge backgrounds and different interest preferences.This paper expand s the research about personalized search. First, in-depth analysis of ontology-based user interest model and changes in user’ s interest, we proposed an update strategy based on the user interest’s attenuation- incentive model, with the update algorithm of heuristic interest in spreading activation model, we implemented the user interest in changing over time. On the basis of other researchers in personalized sorting work, we proposed the reordering algorithm based on user interests and search results.Concrete work done as follows:(1)This paper presents an user interest model based on ontology and proposes an user interest model evolutionary algorithm. O ntolo gy-based user interest model is a collection of the user interest knowledge represented by ontology modeling primitives, is a concept about a field of the class hierarchy. Compared to other representations, the body can be more thorough and meticulous reflecting users’ interests and relationships,which more accurately describes the user’s interest knowledge. In-depth analyzing of user interest changes,based on the characteristics of human oblivion,we proposed forgotten-motivational strategies to heuristic interest in spreading activation model update algorithm, and user interest over time evolving.(2)This paper presents a querying analysis algorithm. First, analyzed the search user behavior and intentions of the user’s classification, elaborated technology related queries, comparing the advantages and disadvantages of various techniques. Finally, by the method of latent semantic analysis, giving the query expansion algorithm, which is a method of statistical analysis of a variety of semantic relationship, this analysis method is designed to improve the recall and precision rates.(3)This paper presents a re-sorting algorithm based on user interest and search results. In order to make the user pages of interest to be the top surface, in calculating the semantic similarity between the user’s interest and the search results,we proposed re-sorting algorithm based on search results and user’s interest.