- [Home]
- [Research achievement]
- [Research achievement detail]
Title | A Note on Multi-topic Document Classification Method Based upon Statistical Decision Theory (in Japanese) |
---|---|
Authors | Yasunari MAEDA 、Hideki YOSHIDA 、Yoshitaka FUJIWARA 、Toshiyasu MATSUSHIMA |
Released Year | 2005 |
Format | Conference |
Category | Knowledge information processing |
Jounal Name | IEICE technical report |
Jounal Page | vol.105, no.665, IT2005-89, pp.147-152 |
Published Year | 2006 |
Published Month | 3 |
Abstract (English) |
In this paper we treat multi-topic document classification problem. In previous researches some theoretical optimality is guaranteed when the number of data for learning is infinite. We propose new multi-topic document classification methods that minimize error rate with reference to the Bayes criterion when the number of data for learning is finite. And we also propose approximate algorithms in order to reduce computational complexity. |
Note (English) |
1 |
Manuscript | |
Presentation |
Involved Papers
- A Bayes Prediction Algorithm for Regression models with Outliers (in Japanese)
- Asymptotic Property on Nonstationary Sources with Piecewise Constant Parameters (in Japanese)
- Asymptotic Property of Universal Lossless Coding for Independent Piecewise Identically Distributed Sources
- A Bayes Prediction Algorithm for Regression models with Outliers (in Japanese)
- A Note on Spelling Correction Methods based upon Statistical Decision Theory
- A Note on Morphological Analysis Methods based on Statistical Decision Theory
- A note on estimation of states in HMM with unknown parameters (in Japanese)
- A Note on Morphological Analysis with Misspellings (in Japanese)
- A Note on Selective - Repeat ARQ with Unknown Block - Error Rate (in Japanese)