Global reoccurrence measure for keyword extraction
- 信息技术－会议论文 
We propose a new feature called Global Reoccurrence Measure for extracting keywords from a document. It expresses the characteristic of keywords that they reoccur frequently in local areas, and have wide coverage in a document. Experiments show that the global reoccurrence measure can improve the performance of keyword extraction. We also propose a new metric called extraction quality for performance evaluation of keyword extraction. ? 2014 IEEE.