基于组合模型的三角模糊数预测模型

首先以北京市1994-2006年年平均最低气温、年平均气温、年平均最高气温构建三角模糊数序列的三个界点,出于数据整体性考虑将三角模糊数序列转换成等量信息的三个指标数序列,然后,对三个指标数序列分别构建了时间序列ARIMA模型,灰色预测GM(1,1)模型以及BP神经网络模型,最后在单一模型预测结果的基础上通过引入诱导有序加权(IOWA)算子,建立了一种集成IOWA算子的ARIMA-GM-BP的三角模糊组合预测模型。通过最终预测结果对比,组合模型预测精度明显要高于各单一模型。

TRIANGULAR FUZZY NUMBER PREDICTION MODELS BASED ON COMBINATION MODEL

Firstly, the three boundary points of the triangular fuzzy number series were constructed based on the annual mean minimum temperature, annual average temperature and annual average maximum temperature in Beijing from 1994 to 2006. Considering the data integrity, the triangular fuzzy number series was transformed into three index series with equal information. Then, the time series ARIMA model, the Grey Prediction GM (1, 1) model and BP neural network model were constructed respectively for the three index series. Finally, based on the prediction results of a single model, we introduced a weighted triangular fuzzy combination forecasting model with ARIMA-GM- operator and IOWA operator by introducing the induced ordered weighted (IOWA) operator. Through the comparison of the final prediction results, the prediction accuracy of the combined model was obviously higher than that of the single model.