Ghanbari, Najme (2015) The GSA Algorithm at Two Methods of Feature Selection and Weighting Features to Improve the Recognition Rate of Persian Handwrite Digits with Fuzzy Classifier. Technical Report. University of Zabol.
|
Text
Technical Report - last version.pdf Download (17kB) | Preview |
Abstract
Introduction: One of the most efficient ways to improve accuracy in a handwritten digits recognition system is selecting optimal features among the entire set of extracted features. In this project, the two methods proposed to increase the recognition rate of Persian handwritten digits. In two methods Gravitational search algorithm (GSA) was used. Binary Gravitational search algorithm (BGSA) proposed for selecting optimal features. On the other hand, one of the aims of this project is to evaluate the increased recognition rate of Persian handwritten digits by using the BGSA. Also, in different proposed way, Persian handwritten recognition rate improved. In this method instead of choosing some of the features by BGSA, one random weight has been assigned to each feature in order to improve recognition rate. Finding the Weight vector by mathematical and statistical computational methods is very difficult. This Weight vector is obtained with real version of GSA (RGSA).
| Item Type: | Monograph (Technical Report) |
|---|---|
| Subjects: | T Technology > TK Electrical engineering. Electronics Nuclear engineering |
| Depositing User: | Mrs najmeh khajeh |
| Date Deposited: | 30 Nov 2021 06:57 |
| Last Modified: | 30 Nov 2021 06:57 |
| URI: | http://eprints.uoz.ac.ir/id/eprint/2902 |
Actions (login required)
 |
View Item |
