线性组合模型及在人口预测中的应用  

Application of Conjugate Gradient Algorithm in Combining Forecasting Mode of the Population

作　　者：孔春旺[1] 童怀水[1] 陈建龙[1] 

机构地区：[1]东华理工大学,江西抚州344000

出　　处：《江西科学》2009年第4期543-547,共5页Jiangxi Science

摘　　要：介绍了线性组合模型并以江西省人口数预测为例,分别利用线性回归预测模型、灰色预测模型、Logistic人口预测模型进行预测,并把这3种预测模型进行最优组合和线性组合,用共轭梯度法性求预测模型的系数。然后对这5种预测模型的结果进行比较,结果表明线性组合预测模型的效果比单一预测模型效果好,也比最优组合预测模型的精度高。The authors present the linear combination forecast model and take the Jiangxi Province population forecast as an example in this article and use the linear regression forecast model, the gray forecast model and the Logistic population forecast model to carry on the forecast separately. And use these three kinds of forecast model to carry on the most superior combination and the linear combination. The coefficient of the linear combination forecast model is solved with the conjugate gradient algorithm. The authors compare the results of these five forecast models. The results show that the result of the linear combination forecast is better than the sole forecast model result and is also higher precision than the most superior combination forecast model.

关 键 词：共轭梯度法 预测模型 人口预测 

分 类 号：O212[理学—概率论与数理统计] F224[理学—数学]