Modeling the warning system of tour sustainable development with gravitational search algorithm support vector regression

Abstract

In this paper, a warning system is constructed using Gravitational Search Algorithm Support Vector Regression (GSA-SVR).The gravitational search algorithm (GSA) is used to optimize the regularization parameter of Support Vector Regression (SVR)and is compared to particle swarm optimization. First, the history data of each index are normalized to (0,1). Then, the weightsof each index are determined by using grey relationship theory, and meanwhile, the degrees of sustainable development of eachyear are calculated. The sustainable development of each year is used as sample to train the SVR. The trained SVR is utilizedas the warning model to predict the degree of sustainable development in future years. The proposed method is applied to theearly warning of tour sustainable development of Qinhuangdao. The simulation results show the effectiveness of the proposedmethod.