Efficient Feature Selection Using Hybrid Slime Mould- Grey Wolf Optimization

Abstract

Feature selection becomes a prominent approach, especially when the records sets
incorporate multiple variables and functions. It is the process of reducing the input data into
an essential model, by disposing the unimportant variables and enhances the accuracy as well
as the performance of type. In this paper Hybrid slime mould- grey wolf optimization
algorithm is proposed for efficient feature selection by incorporating set of rules which could
deal with the classical feature selection short comings. This algorithm is tested over
prominent datasets with higher variety of distinct variables such as Diabetics, Alzheimer,
Heart, Liver, Zoo, Breast Cancer. Four essential characteristics which makes feature selection
is essential are; to simplify the model by way of lowering the range of parameters, subsequent
to lower the training time, to lessen overfilling by using improving generalization, and to
keep away from adding extra dimensionality. The proposed algorithm is compared with the
state-of-the-art techniques Naïve Base (NB), support Vector Machines (SVM), K Nearest
Neighbors (KNN), the best accuracy of the version is the exceptional classifier. Our
experiments show case the comparative examine at distinct views. Furthermore, critical
evaluation metrics Accuracy, Precision, Recall, F-Measure, Time, RMSE, MAE are used to
evaluate the performance. Experimental consequences exhibits that SVM achieves a higher
performance in all test corporations.