CPSO-SVM Based Petroleum Demand Prediction

摘要

Study of oil demand, oil demand uncertainty, leading to its strong non-linear, sudden change characteristic, causes the linear modeling of traditional method and neural network prediction precision is low. In order to accurately forecast demand, presents a chaos particle swarm optimization of support vector machine oil demand forecasting method (CPSO-SVM). The CPSO SVM parameter optimization, and then using SVM to petroleum demand nonlinear variation modeling, finally to 1989～2007 oil demand data for simulation, the results show that, compared with other oil demand forecast algorithm, CPSO-SVM raised oil demand forecast accuracy, as demand for oil to provide a new method for predicting.