Neural Network Learning Enhancement using Island Model based Differential Evolution Algorithm

Abstract

Classification is a machine learning technique used to predict group membership for data instances. To simplify the problem of classification neural networks are being introduced. In this paper, the adaptation of network weights using Island Model based Differential Evolution (IMDE) was proposed as a mechanism to improve the performance of Artificial Neural Network (ANN). Differential Evolution (DE) has been used to determine optimal value for ANN parameters such as learning rate and momentum rate and also for weight optimization. Island model used multiple subpopulations and exchanges the individual to boost the overall performance of the algorithm. In this paper, fully connected topology is being used. This system proposes an island model based differential evolution algorithm to enhance the learning speed of neural network training. The results have revealed that IMDENN has given quite promising results in terms of convergence rate smaller errors compared to other algorithms.