Lévy Mutation in Artificial Bee Colony Algorithm for Gasoline Price Prediction

In this paper, a mutation strategy that is based on Lévy Probabily Distribution is introduced in Artificial Bee Colony algorithm. The purpose is to better exploit promising solutions found by the bees. Such an approach is used to improve the performance of the original ABC in optimizing Least Squares Support Vector Machine hyper parameters. From the conducted experiment, the proposed lvABC shows encouraging results in optimizing parameters of interest. The proposed lvABC-LSSVM has outperformed existing prediction model, Backpropogation Neural Network(BPNN), in predicting gasoline price.