V-Shaped Binary Version of Whale Optimization Algorithm for Feature Selection Problem

Abstract

Feature selection, often known as FS, is often considered to be a component of the larger issue of global optimization. FS is being used to optimize and improve the quality of huge datasets. This is accomplished by selecting prominent features and minimizing duplicate data in order to provide satisfactory classification performance. Feature selection is an approach that is taken with the goal of reducing the complexity of the classification process while simultaneously improving its level of precision. This approach is significant in a variety of fields, including data mining, data processing, and pattern classification. The primary objective here is to devise a more accurate subset of all of the data that takes into account the relevant sample. In order to solve this issue, the BWO-V method, which is short for the binary form of the whale optimization (WO) technique, has been introduced. In order to transform the findings into binary, the BWO-V makes use of a function called the hyperbolic tan function. Validation of the BWO-V algorithm’s performance is carried out on five datasets taken from the repository at UCI. The quantitative and qualitative results both show that using BWO-V helps limit the number of features picked while also optimizing the classification accuracy with significantly less effort.