A Study And Design Of Personalized Meta Search Engine

Today,we are in the age of "Everything Interconnecting".The amount of information on the Internet hasincreased onthe scale of ZB.With the rapid development of social networks,the Internet is exacerbated by the amount of data growth.However,people still feel it difficult to find exactly information needed on the Internet.Both "Information overload" and "Resources isotropic" becomemore serious.To solve these issues,search engines and recommendation technologies have been developed.While facilitatingusers to access network resources efficiently,both technologies have got rapid development with integration trend.In this paper,meta-search engine technology,social networking and recommendation technologyhave been made a combination,mainly study are as following:First,metasearch engine and its personalized progress are studied,the key technologies of the current meta-search engine are analyzed andsummarized.On the basis of the above,a design method and the construction steps of personalized meta-search engine based on recommendation technology are proposed.Second,the present research situation and key technologies of the core part of recommendation---user interest modeling are analyzed and summarized in this paper.Information on social networking platform—Sina microblog are collected as the input of user interest model.Some changes are made on the BIRCH clustering algorithm in order to propose a modeling method of user interest utilizingthe social network information(UIMSNI),the output of UIMSNI is user interest model presented by hierarchical space vector.Third,the scheduling strategy of the underlying search engines and recommendation strategyin personalized meta-search engine are bothdicussed,andboth solutionsbased on UIMSNI are proposed.A combination of social networking,recommendation technology and Meta-Search engineis achieved.Finally,on the basis of above three studies,a personalized meta-search engine(PMSE)is designed and developed.In this paper,the feasibility of the methods is verified by comparing experiments.Experimental results show that the system can provide personalized search service and satisfy the needs of user interest.