Optimization-based feature selection and hybrid machine learning algorithms for big data classification: An investigation

Abstract

For the big data field, optimization algorithms aid in exploring and categorizing the vast data that cannot be dealt with since it is unstructured data. Therefore, extracting (meaningful) data is valuable for various applications. For this goal, many studies were proposed. However, this article presents a strategy for fast classifying massive data with less processing time. The main aim is to obtain data with higher accuracy with fewer mistakes. An appropriate optimization algorithm is selected to classify the data, extract the relevant features, and exclude the unwanted. These steps are conducted according to a study of the data state and the located environment. Additionally, we utilized the smell bear algorithm and Xerus algorithm as a hybrid optimization algorithm based on (hybrid machine learning feature selection), which helps to make the classification process successful and get high-level results. The verification with some of our counterparts demonstrates the superiority of our proposal.