An enhanced extreme learning machine based on ridge regression for regression

Abstract

The extreme learning machine (ELM) is a novel single hidden layer feedforward neural network, which has the superiority in many aspects, especially in the training speed; however, there are still some shortages that restrict the further development of ELM, such as the perturbation and multicollinearity in the linear model. To the adverse effects caused by the perturbation or the multicollinearity, this paper proposes an enhanced ELM based on ridge regression (RR-ELM) for regression, which replaces the least square method to calculate output weights. With an additional adjustment of ridge regression, all the characteristics become even better. Simulative results show that the RR-ELM, compared with ELM, has better stability and generalization performance.