Research On Ranking Method Of Vertical Search Engine

In recent years,the development of the Internet has brought a huge amount of data.The sharp increasing of web data is both an opportunity and a challenge for the search engine.As an important tool for web data retrieval,search engines,especially vertical search engines,have received more and more attention.In this paper,a vertical search engine based on domain ontology is proposed.The vertical search engine mainly consists of some parts,including the search keywords expansion based domain ontology which mine the semantic information of keywords and provides more specialized word set;the improved results ranking algorithm,combining the knowledge of domain ontology to optimize the retrieval results.This paper builds the vertical search engine of Chinese massage field,based on the domain ontology.The main works and research contents of this paper include the following points.First,a new method of keywords expansion is proposed in this paper.The method collects a large number of data in the field,including the electronic documents and the Internet data.The data is applied to train the Word2 vec model.We find the similar words by domain ontology,and filter the similar words by Word2 vec and “Cilin”.Second,improved results ranking algorithm is proposed.It extracts the nodes of domain ontology,and calculates the weight of nodes.Then it counts the nodes in the document and calculates the domain membership.Last the paper optimizations the ranking algorithm by using the domain membership.Third,avertical search engine is designed and implemented based on the domain ontology of Chinese massage field.In the system of vertical search engine,we build the domain ontology of Chinese massage field,implement the keywords expansion and improve results ranking algorithm,based on the Lucene.