基于B型关联度及GIOWA算子的组合预测模型

Combination Forecast Model Based on B-mode Relational Degree and GIOWA Operator

文章首先利用ARIMA（1,1,1）模型、Holt指数平滑模型、M估计值稳健回归等单项预测方法对我国2008—2016年的就业人数进行预测。随后,通过以B型关联度作为目标函数,并引入GIOWA算子,构建了变权组合预测模型对就业人数进行组合预测。研究结果表明,基于B型关联度和GIOWA算子的组合预测模型的预测精度要显著好于单项预测。

This paper firstly carries out a single forecast of China＇s employment population from 2008 to 2016 by using single forecasting methods such as ARIMA（1,1,1） model, Holt exponential smoothing model and M estimated value steady regression model. Then, the paper constructs changeable weight combination forecast model by using B-mode Relational Degree as objective function and bringing in GIOWA operator. Results show that the forecasting precision of combination forecast based on B-mode Relational Degree and GIOWA operator is remarkably higher than that of single forecast.