|
Assessing readability of Thai text using support vector machines |
|---|---|
| รหัสดีโอไอ | |
| Creator | 1. Patcharanut Daowadung 2. Yaw-Huei Chen |
| Title | Assessing readability of Thai text using support vector machines |
| Publisher | Maejo University |
| Publication Year | 2558 |
| Journal Title | Maejo International Journal of Science and Technology |
| Journal Vol. | 9 |
| Journal No. | 3 |
| Page no. | 355 |
| Keyword | Thai readability,term frequency,feature selection,support vector machines |
| ISSN | 1905-7873 |
| Abstract | The readability of a document is a measure of how easily the document can be read and understood. To select appropriate reading materials for children, techniques that can automatically assess readability are required. The objective of this study is to develop a machine-learning-based technique to assess the readability of Thai text. The experimental corpus, which was divided into training data and test data, consisted of articles selected from the textbooks of primary schools in Thailand. Documents in the corpus were first segmented into terms and then represented by feature vectors. Different combinations of feature sets including term frequencies of selected terms, shallow features and language model features were tested in the experiments. Classification and regression models were learned from the training data using support vector machines. Experimental results confirm that the proposed term-selection method can identify effective term frequency features for assessing the readability of Thai text. |
