A Text Classification method based on a Tree Augmented Naive Bayes Network
碩士 === 國立臺灣科技大學 === 資訊管理系 === 100 === Bayesian network models the relationship between response and a set of features. Naive Bayesian approach which assumes that features are independent for each class is usually adopted for its simplicity and practical considerations. However, Naive Bayesian approa...
| Main Authors: | , |
|---|---|
| Other Authors: | |
| Format: | Others |
| Language: | zh-TW |
| Published: |
2012
|
| Online Access: | http://ndltd.ncl.edu.tw/handle/8wht5y |
| Summary: | 碩士 === 國立臺灣科技大學 === 資訊管理系 === 100 === Bayesian network models the relationship between response and a set of features. Naive Bayesian approach which assumes that features are independent for each class is usually adopted for its simplicity and practical considerations. However, Naive Bayesian approach oversimplifies the correlations among features, leading to low classification accuracy in some applications.
Tree Augmented Naive Bayes alleviates the problems caused by the oversimplified assumptions of the Naive Bayes while preserving the affordable computational cost. However, the determination of the root node of the Tree Augmented Naive Bayes network may have substantial impact on the classification accuracy.
We propose a scheme for determining the root node of the Tree Augmented Naive Bayes network for text classification. Empirical results demonstrate that the proposed scheme can achieve high classification accuracy.
|
|---|