Loading...
Please wait, while we are loading the content...
Similar Documents
1 A Comparative Study of Rocchio Classifier Applied to supervised WSD Using Arabic Lexical Samples
| Content Provider | Semantic Scholar |
|---|---|
| Author | Eid, Soha M. Al-Said, Almoataz B. Wanas, Nayer M. Rashwan, Mohsen A. A. Hegazy, Nadia H. |
| Copyright Year | 2015 |
| Abstract | This work studies the possibilities of using the Rocchio classifier to solve the Word Sense Disambiguation problem in a supervised manner through the usage of the lexical samples of five Arabic words. The performance of the Rocchio classifier is compared to three supervised machine learning algorithms, namely the Most Frequent Sense (MFS), Naïve Bayesian Classifier (NBC) and the Support Vector Machine (SVM) representing the baseline and stateof-the-art algorithms for WSD. Results indicate that the Rocchio classifier outperforms the other classification approaches by reducing the error by over 14% compared to the best performing NBC due to its superior ability in feature selection. |
| File Format | PDF HTM / HTML |
| Alternate Webpage(s) | https://scholar.cu.edu.eg/?q=moataz/files/wsd_i.pdf |
| Language | English |
| Access Restriction | Open |
| Content Type | Text |
| Resource Type | Article |
