Word similarity algorithm based on multi-features. Issue 5 (4th July 2018)

• Record Type: Journal Article
• Authors: Guo, Chang-Jin; Pan, Feng; Zuo, Yi
• Is Part Of: Journal of interdisciplinary mathematics. Volume 21:Issue 5(2018)
• Page Start: 1067
• Page End: 1072
• Publication Date: 2018-07-04
• Subjects: Word similarity -- Multi-features fusion -- Natural language processing; Mathematics -- Periodicals; Mathematics; Periodicals; 510.5
• DOI: 10.1080/09720502.2018.1493031
• Languages: English
• ISSNs: 0972-0502

Abstract

Abstract: Word similarity calculation is one of key problems in natural language processing, In this paper, a multi-features fusion word similarity algorithm was proposed based on features of knowledge base and corpus statistics, of which the correlation coefficient to the human annotated similarity test dataset M&C is 0.872. Results show that such method can effectively calculate the similarity between words.