A Review on Feature Selection Approaches and Few Prominent Research Gaps for Image Classification

This paper presents approaches used for image classification. Feature selection is an important task which reduces time complexity to a greater extent and yields better results. It focuses on selecting relevant features and excluding irrelevant features. Various methods have been studied for formulating this work and significant research gaps have been marked for the authors to work. Several research gaps and papers containing Evolutionary computation (EC), Information Gain binary Butterfly Optimization Algorithm (IG-bBOA), Multi-objective genetic algorithm (MOGA), mRMJR-KCCA, MI based hybrid model, MI-SNN and MI-DNN are studied for formalizing this work. Feature selection problem for early time series classification and ethno botanical study on medicinal plants are also taken into consideration. Research gaps are framed based on extensive studies of literature. Research answers from literature and future directions are identified.