A key component extraction method based on HMM and dependency parsing

Abstract

Increasing attention has been paid for POI (Point of Interest) data query for travel information service. The correct extraction of key components in question is crucial for improving the accuracy of query results. The paper proposes a key component extraction method based on HMM (Hidden Markov Model) and dependency parsing. Firstly, the sentence pattern classifier is established by HMM. And then, questions are classified by classifier. Finally, combination of sentence pattern's structure, the four key components are extracted by dependency parsing. The results show that the F1-Measure is 0.83, which well proves the effectiveness of the method.