The Intelligent Full Text Retrieval System Based On Topic Sorting And Recommendation

With the rapid development of information technology in the Internet,the blowout of various portal webs and the explosive growth of network data have made it increasingly urgent for people to obtain knowledge what they need from massive information.For individuals,how to quickly and accurately find the destination page from the vast amount of information is the key.For each web site with a large number of pages,how to quickly build an accurate and personalized search system becomes a top priority.Under the above demand,this paper proposes an intelligent full-text retrieval search engine based on topic ordering and recommendation.The main work contents of this paper are as follows:First,analysis the research background and the meaning of the system,and introduce the development of full-text search engine and sorting technology and its present situation in China and abroad.At the same time,this paper analysis the specific architecture and recommended model modeling of a full-text search engine system.Clarifies the business requirements and processes,and proposes the specific process of building this intelligent full-text search engine.Secondly,this paper divides the system into four layers of module structure.Firstly,it sorts the retrieval results,uses distance frequency correlation algorithm and LDA theme model to match the content.Uses Pagerank algorithm to calculate the importance of links,and uses BP neural network and user records feedback on the sort of learning results and optimization.At last,use the weighted of several algorithms to make a comprehensive ranking,which makes the retrieval result more reasonable.At the same time,it explores the theory of personalized recommendation.Based on the characteristics of the topic analysis and retrieval system,a Hybrid Recommendation algorithm based on Topic(HRT)is proposed to explore user's main preference topic and potential preference topic,and adopt a hybrid recommendationcombining two ways,which can also effectively solve the problem that the recommended algorithm cold-started in the search engine application.Finally,the system design and implementation,mainly using Python to achieve each module,while designing an error design and provide multiple sets of programs for the stable operation of the system.Analysis the results of the system functions and metrics in detail,and the use of black-box test methods and LoadRunner load test tools to test the system.The results show that the system basically reached the needs of intelligent retrieval.In addition,the system's modular design and flexible optimization algorithm to provide users with a reasonable page ranking,and on the basis of the actual function to ensure the stability of the system,accuracy and intelligence and high scalability to help users from the mass the most convenient and fast way to find the information they need.Effectively solves the problem of user intelligent information retrieval and the quick and personalized deployment of website sites.