天然气消费量的偏最小二乘支持向量机预测

结合偏最小二乘法和支持向量机的优缺点,提出基于偏最小二乘支持向量机的天然气消费量预测模型。首先,利用偏最小二乘法确定影响天然气消费量的新综合变量,建立以新综合变量为输入,天然气消费量为输出的支持向量机模型,对天然气消费量进行了预测;然后,与多元回归、偏最小二乘回归、普通支持向量机做误差检验比较,验证该方法的可行性与正确性。结果表明,此天然气消费量预测模型具有较高的精确度和应用价值。

Prediction of Natural Gas Consumption Based on Partial Least-squares Support Vector Machines Regression

Combined with the advantages of partial least squares and support vector machine (SVM) ,partial least squares support vector machine regression method of predicting natural gas consumption was brought forward .First use partial least squares method to extract the new synthetic variables that affect natural gas consumption. Second, use support vector machines regression to establish the nonlinear predicting model which the new synthetic variables were used as the input and the natural gas consumption as the output. Finally, compare the error of PLS-SVM with PLS,SVM and multiple regression to verify the feasibility and correctness of the method. The results show that the method is correct and feasible.