Extreme Learning Machine for Regression and Classification Using L 1-Norm and L 2-Norm

Abstract

Extreme learning machine (ELM) has been studied extensively in recent years. It is a very simple machine learning algorithm which can achieve a good generalization performance with extremely fast speed. Thus, it has practical significance for Big Data analysis. Normally, it is implemented under the empirical risk minimization scheme and it may tend to generate a large-scale and over-fitting model. In this paper, an ELM model based on L 1-norm and L 2-norm regularizations is proposed to deal with regression and multiple class classification problems in a unified framework, and it can reduce the complexity of the network and prevent over-fitting. We test the proposed algorithm on eight benchmark data sets. Simulation results have shown that the proposed algorithm outperforms the original ELM and other advanced ELM algorithm in terms of prediction accuracy and stability.