Learning Semantic Lexicons using Graph Mutual Reinforcement based Bootstrapping

Bootstrapping has been received a amount of attentions in many fields and achieved good results. While semantic lexicons also have been proved to be useful for many natural language processing tasks. This paper presents an approach to learn semantic lexicons using a new bootstrapping method which is based on Graph Mutual Reinforcement. The approach uses only unlabeled data and a few of seed words to learn new words for each semantic category. Different with other bootstrapping methods, we use Graph Mutual Reinforcement based Bootstrapping to sort the candidate words and patterns. Experimental results show that GMR-Bootstrapping outperforms the state-of-the-art algorithms both in in-domain data and out-domain data. Furthermore it is also shows that the result was depended on not only the size of the corpus, but also the quality.