Keyphrases are very useful and significant for information retrieval, automatic summarizing, text clustering, etc. KEA is a traditional and classical algorithm in keyphrase automatic extraction. But it is mainly based on the statistical information without considering the semantic information. In this paper, we propose a method which combine semantic information with KEA by constructing lexical chain that based on Reget's thesaurus. In our method, the semantic similarity between terms is used to construct the lexical chain, and then we use the length of the chain as a feature to build the extraction model. The experiment result shows that the performance of the system has a big improvement compare with the KEA.
展开▼
