| Supervised feature selection via information gain, maximum projection and minimum redundancy |
| کد مقاله : 1079-SLAA10 |
| نویسندگان |
|
Mahdi Eftekhari1، Farid Saberi-Movahed *2، Adel Mehrpooya1 1Department of Computer Engineering, Shahid Bahonar University of Kerman, Kerman, Iran 2Department of Applied Mathematics, Faculty of Sciences and Modern Technologies, Graduate University of Advanced Technology, Kerman, Iran |
| چکیده مقاله |
| Feature selection problem is an important issue in both data clustering and data classification. This paper introduces a supervised framework for the task of feature selection. The proposed method is built based on applying the information gain method into the framework of maximizing relevancy, and aims to reduce the redundancy between the selected features by using the idea of maximum projection and minimum redundancy. Several experimental results on seven well-known microarray datasets demonstrate the promising performance of the proposed method over some state-of-the-art methods in this area. |
| کلیدواژه ها |
| Machine learning, Supervised feature selection, Information gain, Maximum projection, Minimum redundancy |
| وضعیت: پذیرفته شده مشروط برای ارائه شفاهی |