Presentation | 2001/7/9 Estimating reliability of the rules in decision lists using Bayesian learning Yoshimasa Tsuruoka, Takashi Chikayama, |
---|---|
PDF Download Page | PDF download Page Link |
Abstract(in Japanese) | (See Japanese page) |
Abstract(in English) | The decision list algorithm is oneof the most successful algorithms for classification problems in natural language processing. We propose a method based on Bayesian learning to calculate the reliability of contextual evidences in decision lists. The method also gives well-founded smoothing and better use of prior information of each type of contextual evidence. We evaluate these proposed methods on Japanese word sense disambiguation problems. The results show improved accuracy close to one expected from the Bayesian theory. |
Keyword(in Japanese) | (See Japanese page) |
Keyword(in English) | Decision list / Bayesian learning / Prior distribution |
Paper # | NLC2001-21 |
Date of Issue |
Conference Information | |
Committee | NLC |
---|---|
Conference Date | 2001/7/9(1days) |
Place (in Japanese) | (See Japanese page) |
Place (in English) | |
Topics (in Japanese) | (See Japanese page) |
Topics (in English) | |
Chair | |
Vice Chair | |
Secretary | |
Assistant |
Paper Information | |
Registration To | Natural Language Understanding and Models of Communication (NLC) |
---|---|
Language | JPN |
Title (in Japanese) | (See Japanese page) |
Sub Title (in Japanese) | (See Japanese page) |
Title (in English) | Estimating reliability of the rules in decision lists using Bayesian learning |
Sub Title (in English) | |
Keyword(1) | Decision list |
Keyword(2) | Bayesian learning |
Keyword(3) | Prior distribution |
1st Author's Name | Yoshimasa Tsuruoka |
1st Author's Affiliation | School of Engineering, The University of Tokyo() |
2nd Author's Name | Takashi Chikayama |
2nd Author's Affiliation | School of Frontier Sciences, The University of Tokyo |
Date | 2001/7/9 |
Paper # | NLC2001-21 |
Volume (vol) | vol.101 |
Number (no) | 189 |
Page | pp.pp.- |
#Pages | 7 |
Date of Issue |