Open AccessResearch on Query Term Expansion based on RankSVM and LDA Model
Author(s) -
Peng Lin,
Guoquan Lu
Publication year - 2020
Publication title -
journal of physics. conference series
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 0.21
H-Index - 85
eISSN - 1742-6596
pISSN - 1742-6588
DOI - 10.1088/1742-6596/1684/1/012050
Subject(s) - query expansion , computer science , query optimization , term (time) , sargable , sorting , web query classification , set (abstract data type) , result set , relevance feedback , relevance (law) , information retrieval , web search query , data mining , extension (predicate logic) , search engine , artificial intelligence , algorithm , image retrieval , physics , quantum mechanics , political science , law , image (mathematics) , programming language
The lack of semantic association will lead to poor performance of retrieval system. This paper proposed a query terms extension method based on RankSVM and LDA model. Firstly, with the aid of RankSVM’s excellent performance of sorting, high-quality initial retrieved results will be obtained. Then utilizing the LDA model to select the documents related to query term and to generate topic model for sequenced initial retrieval results.in the post-processing step, using threshold to select the highest probability among each topic and regards them as query terms set. Experiments show that the query terms expansion method proposed in this paper is superior to the current query expansion methods based on pseudo-relevance feedback in the five evaluation index, therefore this is a feasible method.