A novel feature selection algorithm using multi-objective improved honey badger algorithm and strength pareto evolutionary algorithm-II

Abstract EN

Feature selection is an important task in classification that removes redundant or irrelevant features from the dataset.

Many researchers favor a multi-objective feature selection approach.

However, these approaches fail to maintain high classification accuracy while removing redundancy in the features.

In this work, a wrapper-based feature selection technique is proposed using a hybrid of the Multi-Objective Honey Badger Algorithm (MOHBA) and the Strength Pareto Evolutionary Algorithm-II, called MOHBSP2, to balance classification accuracy and redundancy removal.

Classification accuracy improvement and the removal of redundant features are considered the multi-objective optimization functions of the proposed multi-objective feature selection technique.

The Levy flight algorithm is used to initialize the population and enhance the exploration and exploitation of MOHBA.

The regularized Extreme Learning Machine (ELM) is used to classify the selected features.

To evaluate the performance of the proposed feature selection technique, 18 benchmark datasets are used and results are compared with the four well-known multi-objective feature selection techniques in terms of accuracy, hamming loss, ranking loss, mean value, standard deviation, feature length, and training time.

The proposed approach achieved a maximum accuracy of 99% with the maximum value of selected features as 80.

The minimum value of hamming loss, ranking loss, mean value, and standard deviation value achieved by the proposed approach are 0.0092, 0.0003, 0.018, and 0.001, respectively.

The experimental results show that the proposed approach can improve classification accuracy and remove redundancy in large datasets.