Parameter Optimization of Kernel Extreme Learning Machine Using Artificial Bee Colony Algorithm and Its Application for Disease Classification

Abstract

Machine learn methods have been widely used for classification and diagnosis of diseases for increasing its accuracy and efficiency. The kernel extreme learning machine is being increasingly used algorithm to training single layer forward neural network as that this network is given the weights between input and hidden layers, and the bias parameter of each hidden node. In order to obtain more stable and accurate model, an artificial bee colony algorithm is used to pre-train parameters of kernel parameter and penalty parameter. weight and bias. In this paper, an artificial bee colony based kernel extreme learning machine is proposed to classify medical datasets. This proposed method is called ABC-KELM. In experiments, we use two benchmark datasets that are Breast cancer and Parkinson disease from the UCI repository to evaluate the effectiveness and classification accuracy. The experimental results reveal that the ABC-KELM can obtain satisfactory classification results.