- [Home]
- [Research achievement]
- [Research achievement detail]
Title | A Note on Document Classification Methods with Small Training Data (in Japanese) |
---|---|
Authors | Yasunari Maeda 、Hideki Yoshida 、Masakiyo Suzuki 、Toshiyasu Matsushima |
Released Year | 2011 |
Format | Journal |
Category | Knowledge information processing |
Jounal Name | IEEJ Transactions on Electronics, Information and Systems |
Jounal Page | vol.131, no.8, pp.1459-1466 |
Published Year | 2011 |
Published Month | 8 |
Abstract (English) |
Document classification is one of important topics in the field of NLP(Natural Language Processing). In the previous research a document classification method has been proposed which minimizes an error rate with reference to a Bayes criterion. But when the number of documents in training data is small, the accuracy of the previous method is low. So in this research we propose a new document classification method using estimating data in order to estimate prior distributions, which is based on the previous method. When the training data is small the accuracy of the proposed method is higher than the accuracy of the previous method. But when the training data is big the accuracy of the proposed method is lower than the accuracy of the previous method. So in this research we also propose another document classification method whose accuracy is higher than the accuracy of the previous method when the training data is small, and is almost the same as the accuracy of the previous method when the training data is big. |
Note (English) |
3 |
Manuscript | |
Presentation |
Involved Papers
- A Note on Mixed Level Experimental Designs Using Augmented Orthogonal Arrays (in Japanese)
- 半教師付き学習における一致性を満たすゆう度方程式の解に基づく予測の漸近評価
- A CLASS OF NOISELESS CODES DESIGNED BY DECISION THEORY
- 相互情報量最大に基準を置くユーザインタフェースの効率化
- パターンごと・ステージごとに事後確率のしきい値をおくストッピングルール
- Inductive Inference and Description Length (in Japanese)
- 信頼性を考慮した推論について
- On Knowledge Representation and Reasoning System with a Range of Certainty Factors (in Japanese)
- 特集にあたって
- On Statistical Model Selection based on Bayes Decision Theory (in Japanese)
- On Complexity of Decoding beyond the BCH Bound Using Berlekamp-Massey Algorithm (in Japanese)
- On Decoding Methods beyond the BCH Bound and their Applications to Soft-Decision Decoding (in Japanese)
- Soft-Decision Decoding Using Decoding Method beyond the BCH Bound for Binary BCH Codes (in Japanese)
- A Study on Difference of Codelengths between Codes Based on MDL Principle and Bayes Codes for Given Prior Distributions (in Japanese)
- A Decision Feedback Scheme Using List Decoding for Tree Codes (in Japanese)
- Asymptotic Normality of Extended Posterior Density Functions with Loss Functions (in Japanese)
- A New Decoding Algorithm Using Likelihood Ratio Testing for Tree Codes (in Japanese)
- On Error Exponents for Variable Size List Decoder Using the Viterbi Algorithm with Likelihood Ratio Testing (in Japanese)
- On the Interleaver Design Method for Block Turbo Codes and Its Minimum Distance (in Japanese)
- あいまいな命題を含む推論モデルに関する一考察