Learning Semantic Lexicons Using Graph Mutual Reinforcement Based Bootstrapping

This paper presents a method to learn semantic lexicons using a new bootstrapping method based on graph mutual reinforcement (GMR). The approach uses only unlabeled data and a few seed words to learn new words for each semantic category. Different from other bootstrapping methods, we use GMR-based bootstrapping to sort the candidate words and patterns. Experimental results show that the GMR-based bootstrapping approach outperforms the existing algorithms both in in-domain data and out-domain data. Furthermore, it shows that the result depends on not only the size of the corpus but also the quality.