The Vertical Search Engine Page Ranking Algorithm Based On Domain Ontology Research

With the dramatic increase in network information resources, the importance of search engine is growing. The page sorting algorithm is a key part of the search engine. How to effectively find the information you need is critical, a good search engine can greatly save time for users to find information. Search engine contains several components, the accuracy of the sorting results of web pages directly determines the performance of search engine and user experience. There are many page sorting algorithms in areas of information retrieval, algorithm which based hyperlink analysis is more extensive. Through the study of related technology the vertical search engine’s working principle and architecture, ontology, etc. on this basis, to construct model based on e-commerce domain ontology, semantic factors, the improvement of sorting algorithm was carried out in-depth study. The main research contents concentrate on the following aspects:1. By studying the existing page ranking algorithm, analyzes the existing deficiency, and introduces the evaluation standard of web page. In view of the returned results correlation, user experience, response time and so on, the improved strategy based on PageRank algorithm is proposed, to apply semantic factors in the vertical search sort algorithm, so as to improve the accuracy of search results.2. The analysis of the ontology technology, build rules, the modeling of metalinguistic and classification, build the e-commerce ontology, and will apply WordNet semantic relationships dictionary in vertical search engine sorting algorithm. On this basis, propose concept semantic similarity computation based on e-commerce. Through the programming to realize the word network, namely, given a word can find out all the synonyms for the word, and the similarity of the improved algorithm is verified through the experiment under the environment of Chinese and English semantic similarity calculation results.3. The vertical search engine system based on domain ontology is realized, the improved PageRank algorithm will be used in this system, including the information acquisition module, the Lucene index module, ontology construction and management module and query expansion and the results display module. Finally, using Loadrunner performance test tool in the throughput, the average response time, Hits/second three aspects verifies the performance of the system, and finally it is concluded that higher value degree page, and thus meet the needs of users.