Hyper parameters Optimization of Support Vector Regression based on a Chaotic Pigeon-Inspired Optimization Algorithm

Abstract

Support vector regression (SVR) as a data mining tool has been applied in several real problems. However, it is usually needed to tune manually the hyperparameter. Meta-heuristic algorithms have been used as hyperparameters tuning procedure. In this paper, chaotic pigeon-inspired optimization algorithm is proposed to enhance the exploration and exploitation capability of the pigeon-inspired optimization algorithm.The results, which were applied to four of the datasets, show that the proposed algorithm has the possibility to obtain results better than the cross-validation method through prediction values ??and time spent, as well as the efficiency of the proposed algorithm in improving prediction and computational time spent in processing, compared to algorithms inspired by nature.