Application of SVM Combined with Mackov Chain for Inventory Prediction in Supply Chain

The aim of this paper is to predict the inventory of the relevant upstream enterprises in supply chain. The support vector machine, a novel artificial intelligence-based method developed from statistical learning theory, is adopted herein to establish a short-term stage forecasting model. However, take the fact into account that demand signal is affected by variant random factors and behaves big uncertainty, the predicted accuracy of SVM is not approving when the data show great randomness. It is obligatory that we present Markov chain to improve the predicted accuracy of SVM. This combined model takes advantage of the high predictable power of SVM model and at the same time take advantage of the prediction power of Markov chain modeling on the discrete states based on the SVM modeling residual sequence. Then we use the statistical data of the output of the gasoline of China from Feb-06 to Dec-07 for a validation of the effectiveness of the above model.