Presentation | 2010-05-14 A novel method for minimum entropy clustering using competitive learning Toshio UCHIYAMA, Takeharu EDA, Katsuji BESSHO, Tadasu UCHIYAMA, Katsuyoshi TANABE, |
---|---|
PDF Download Page | PDF download Page Link |
Abstract(in Japanese) | (See Japanese page) |
Abstract(in English) | We propose a novel method(algorithm), which uses competitive learning, for clustering based on minimizing average entropy in clusters(Minimum entropy criterion). In each iteration, competitive learning usually decides the winner of models for an input data and then updates the model of the winner. When deciding the winner, minimum entropy criterion requires calculation of the cross-entropy against the models. However, it is not easy to manage values of stochastic variables in models with avoiding zero frequency problem, because competitive learning updates models so many times. Therefore, we present the idea, which updates models before deciding a winner, to avoid the zero frequency problem without any bad influence. We also propose a new method to apply the clustering technique for improving text classification accuracy. The effectiveness of our methods is shown by experiments. |
Keyword(in Japanese) | (See Japanese page) |
Keyword(in English) | Clustering / KL-divergence / Minimum entropy criterion / Competitive learning / Naive Bayes classifier |
Paper # | IE2010-39,PRMU2010-27,MI2010-27 |
Date of Issue |
Conference Information | |
Committee | MI |
---|---|
Conference Date | 2010/5/6(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 | Medical Imaging (MI) |
---|---|
Language | JPN |
Title (in Japanese) | (See Japanese page) |
Sub Title (in Japanese) | (See Japanese page) |
Title (in English) | A novel method for minimum entropy clustering using competitive learning |
Sub Title (in English) | |
Keyword(1) | Clustering |
Keyword(2) | KL-divergence |
Keyword(3) | Minimum entropy criterion |
Keyword(4) | Competitive learning |
Keyword(5) | Naive Bayes classifier |
1st Author's Name | Toshio UCHIYAMA |
1st Author's Affiliation | NTT Cyber Solutions Laboratories, NTT CORPORATION() |
2nd Author's Name | Takeharu EDA |
2nd Author's Affiliation | NTT Cyber Solutions Laboratories, NTT CORPORATION |
3rd Author's Name | Katsuji BESSHO |
3rd Author's Affiliation | NTT Cyber Solutions Laboratories, NTT CORPORATION |
4th Author's Name | Tadasu UCHIYAMA |
4th Author's Affiliation | NTT Cyber Solutions Laboratories, NTT CORPORATION |
5th Author's Name | Katsuyoshi TANABE |
5th Author's Affiliation | NTT Cyber Solutions Laboratories, NTT CORPORATION |
Date | 2010-05-14 |
Paper # | IE2010-39,PRMU2010-27,MI2010-27 |
Volume (vol) | vol.110 |
Number (no) | 28 |
Page | pp.pp.- |
#Pages | 6 |
Date of Issue |