A Method for Forecasting Regional Logistics Demand Based on Support Vector Machine and Its Application

This paper presents a method for forecasting regional logistics demand based on support vector machine. The characteristics of this model are concluded as follows: First, using support vector regression with linear kernel function to analyze influences of the input influential factors to the regional logistics demand and determine the key influential factors. Second, constructing a fusion kernel based on wavelet analysis to replace the Gauss kernel. This study applies the proposed method to predict the regional logistics demand of Shanghai during the period 1978 to 2005. Experiments show that this method has very good prediction accuracy.