Moth-Flame Optimization Algorithm for Feature Selection: A Review and Future Trends

Description

  • Qasem Al-Tashi, Seyedali Mirjalili, Jia Wu, Said Jadid Abdulkadir, Tareq M Shami, Nima Khodadadi, Alawi Alqushaibi

  • Download Paper

  • 20 September 2022

Feature selection assists with finding an optimal set of features as a preprocessing step before using datasets in Machine Learning algorithms. As it selects features, it is considered as a binary problem [15]. MFO has been designed to solve problems with continuous values, so we need to change it to solve binary problems. Transfer functions map continuous values of algorithm to binaries, which is what we are looking for when using MFO to solve binary problems. Four S-Shaped as well as four V-Shaped transfer functions are presented in this review. The mathematical equations of the eight transfer functions are used in this chapter, and it should be noted that each transfer function provides a different performance. Moreover, only some transfer functions are utilized on MFO which motivated researchers in the literature to apply the rest of the transfer functions on MFO and evaluate their performance.

Codes

Leave a Comment