A New Term Significance Weighting Approach

Authors: Jin Zhang¹; Tien Nguyen²

Source: Journal of Intelligent Information Systems, Volume 24, Number 1, January 2005 , pp. 61-85(25)

Publisher: Springer

Buy & download fulltext article:

Price: $47.00 plus tax (Refund Policy)

Abstract:

The authors present a new term significance measure that integrates term frequency retrieval characteristics, term frequency, document collection characteristics, and both the term depth and width distribution characteristics. A new concept, the term depth distribution, is introduced and its impact on the term significance is analyzed. The authors address the features of the new term significance measure from the angles of the impact of the variables (parameters) on it and the iso-significance contour analyses. An experimental study was conducted to compare the newly developed approach with two other popular approaches from the perspectives of both efficiency and effectiveness. The results show that the newly developed approach achieves satisfactory performance. Issues for further research on this topic are suggested.

Keywords: term significance; automatic term weighting; term weighting evaluation

Document Type: Research article

DOI: http://dx.doi.org/10.1007/s10844-005-0267-y

Affiliations: 1: University of Wisconsin-Milwaukee, Bolton Hall 532, P.O. Box 413, Milwaukee, Wisconsin, 53201, Email: jzhang@csd.uwm.edu 2: University of Wisconsin-Milwaukee, 3200 North Cramer Ave, EMS Building, Milwaukee, Wisconsin, 53201, Email: tien@cs.uwm.edu

Publication date: 2005-01-01