Concept Extraction Based on Hybrid Approach Combined with Semantic Analysis

Abstract

Abstract. This paper proposed a hybrid approach for concept extraction, which combined linguistic, statistical and semantic analyses. Similar to traditional methods, linguistic and statistical analyses were reserved; this paper innovatively adopted semantic analysis to analyze the semantic relationships of concepts in order to get a fine grained result. Another creative work is the combination of several statistical approaches from different perspective to filter the candidate concepts. Experiments had shown a promising result.