Research on Customers Demand Forecasting for E-business Web Site Based on LS-SVM

This paper introduces a novel customers´ demand forecasting model based on least squares support vector machines (LS-SVM) for e-business enterprises. Firstly, the paper presents actual state of e-business, and discusses some factors that block e-business advance in China. Then, some common techniques used for forecasting are briefly reviewed together with their shortcomings respectively. To solve these disadvantages, the paper reviews the fundamental theory of least squares support vector machines for regression, and analyses some merits of the theory. At last, based on the theory, the paper proposes a forecasting model to forecast pure water demand in a week for an e-business website. Compared with linear neural network predictor, RBF neural network predictor and BP neural network predictor, the LS-SVM forecasting model shows outstanding performance in simulation and practical results.