Automatic Semantic Role Labeling for Chinese Comparative Sentences Based on Hybrid Patterns

Automatic semantic role labeling of comparative sentences contributes to opinion mining and information recommendation and so on. In this paper, an effective method of semantic role labeling based on hybrid comparative patterns is proposed for Chinese comparative sentences. In this method, the original hybrid comparative patterns are constructed according to the syntactic structures of comparative sentences. Then they are generalized to improve the accuracy and coverage rate of the labeling results. A two-level algorithm is designed for comparative entities labeling and comparative features labeling respectively. The experimental results indicate the effectiveness of the proposed method.