Effective Classification Using Artificial Bee Colony Based Feature Selector

Authors

  • S.H. Hassan Faculty of Computer Science and Information Technology, Universiti Malaysia Sarawak, 94300, Kota Samarahan Sarawak.
  • N. Yusup Faculty of Computer Science and Information Technology, Universiti Malaysia Sarawak, 94300, Kota Samarahan Sarawak.
  • D.N.F. Awang Iskandar Faculty of Computer Science and Information Technology, Universiti Malaysia Sarawak, 94300, Kota Samarahan Sarawak.

Keywords:

Artificial Bee Colony, Image Processing, Classification Technique, Feature Selection,

Abstract

Pepper classification is an important step in measuring the quality of the pepper. As of today, Malaysia Pepper Board (MPB) performed its classification task by doing a semi-automated process; where a commercial colour sorter is used to differentiate and categorized the peppers into respective groups. However, it is proven to be a challenging and time consuming process This paper introduces an effective pepper classification by applying a feature selection method in Artificial Bee Colony (ABC) algorithm. ABC algorithm is a metaheuristic algorithm inspired by the collective behavior of bees, which has been used successfully to solve complex problems of optimization. From this study, the proposed ABC algorithm that incorporates feature selection method resulted an accuracy of 92% tested with a small sample size. Meanwhile, 89.7% of accuracy was obtained with a larger sample size and a set of Red Mean; Green Median; Red Standard Deviation; Solidity Ratio; and Contrast were chosen as the best optimal features' subset.

Downloads

Download data is not yet available.

Downloads

Published

2017-12-07

How to Cite

Hassan, S., Yusup, N., & Awang Iskandar, D. (2017). Effective Classification Using Artificial Bee Colony Based Feature Selector. Journal of Telecommunication, Electronic and Computer Engineering (JTEC), 9(3-11), 115–120. Retrieved from https://jtec.utem.edu.my/jtec/article/view/3194