Abstract
Term weighting is a vital task in nature language processing domain. It is important means for text semantic representation. As an essential part, the term global weight is also obtained by statistically methods. To achieve relatively logical term global weight for every word in lexicon, this paper compares several popular term weighting methods and proposes a novel combining method of term global weighting through theoretical analysis and experiments. Experimental results revealed that the modified expected cross entropy method yielded better performance than the others in our application.
| Original language | English |
|---|---|
| Pages (from-to) | 315-318 |
| Number of pages | 4 |
| Journal | Harbin Gongye Daxue Xuebao/Journal of Harbin Institute of Technology |
| Volume | 43 |
| Issue number | SUPPL. 1 |
| State | Published - Mar 2011 |
| Externally published | Yes |
Keywords
- Information Retrieval
- Semantic similarity computation
- Term global weighting
Fingerprint
Dive into the research topics of 'A novel term global weighting method'. Together they form a unique fingerprint.Cite this
- APA
- Author
- BIBTEX
- Harvard
- Standard
- RIS
- Vancouver