基于预测模型的中国原木进口数据预测研究

Prediction of Log Import Data in China Using GM(1,1)and LSM Models

为了准确地预测未来原木进口量,利用已知中国进口原木数据,使用灰色模型GM(1,1)模型以及最小二程法多项式拟合的经验公式对1993-2013年的国内原木进口数据进行预测,通过预测值与实际值的对比,分析两组预测模型的准确性。结果表明:基于GM(1,1)模型的预测结果,在RMSE、MAE、MAPE三个指标上均优于基于最小二乘法函数拟合的预测方法。利用GM(1,1)模型预测2014和2015年国内原木进口量分别是4 170.58万和4 358.99万m3。

Through the existing data ,the grey model GM （1 ,1） prediction model and minimum two‐way method of polynomial fitting empirical formulas ,domestic timber import data of 1993‐2013 years was predicted ,through the predicted and actual values were comparied ,two kinds of forecasting model were analyzed .The results showed that based on the GM （1 ,1） model prediction in RMSE ,MAE and MAPE were better than the predic‐tion method based on least squares fitting .The use of GM （1 ,1） model to forecast in 2014 and 2015 domestic log import volume was 41 .705 8 million and 43 .589 9 million cubic meter respectively .