ELM combined with hybrid feature selection for classification

Abstract

The purpose of data classification is to assign new data objects to a correct category based on their attributes. But the complexity of data, high dimensional feature space and low quality of feature selection become the main problem for data classification process. In order to improve the classification performance and reduce the data dimension, effective feature selection is becoming increasingly important. In this paper, an improved F -score and heuristic search strategy are used as feature selection methods, extreme learning machine (ELM) is used to evaluate the selected feature subsets and used to achieve effective feature selection. Five-fold cross-validation is used in experiments, and experimental data sets are taken from the UCI machine learning repository. The experimental results show that the feature selection method based on improved F -score and heuristic search not only make the input of classifier have fewer features, but also lead the classifier to have a higher classification accuracy, less time consuming, and a good generalization performance.