| 引用本文: | 王冠智,粟晓玲,张特,等.基于DWT-WFGM(1,1)-ARMA组合模型的农业用水量预测[J].灌溉排水学报,0,():-. |
| Wang Guanzhi,Su Xiaoling,Zhang Te,et al.基于DWT-WFGM(1,1)-ARMA组合模型的农业用水量预测[J].灌溉排水学报,0,():-. |
|
| 摘要: |
| 农业用水量预测对于区域水资源规划与管理具有重要意义。【目的】针对农业用水量序列的振荡性以及传统模型预测结果输出单一的问题，提出一种新的组合预测模型DWT-WFGM(1,1)-ARMA。【方法】通过离散小波变换将原始用水量序列分解为近似序列和细节序列，并分别采用自回归滑动平均模型和分数阶灰色模型预测细节序列和近似序列，并结合加权马尔可夫链对近似序列进行误差修正，将不同成分序列的预测结果进行线性叠加得到农业用水量的预测值及预测区间。【结果】利用该模型分别对陕西省和内蒙古省的农业用水量进行预测，并与灰色模型GM(1,1)、DWT-GM(1,1)-ARMA模型和DWT-FGM(1,1)-ARMA模型对比分析。DWT-WFGM(1,1)-ARMA模型在陕西省和内蒙古省的评价指标平均绝对百分比误差分别为1.25%和1.01%，精度明显高于其它模型，且预测区间为研究区未来时期的农业用水量提供了合理的波动范围，具有一定的实用性。【结论】综上所述，本文构建的模型能够有效提高农业用水量预测的精度，同时预测区间的提出可以为区域农业用水量预测提供更加可靠依据。 |
| 关键词: 农业用水；分数阶灰色模型；加权马尔可夫链；离散小波变换；预测区间 |
| DOI： |
| 分类号:TV213.4 |
| 基金项目:国家自然科学基金项目（52079111），“十三五”国家重点研发计划项目（2016YFC0401306） |
|
| Combined Model for Prediction of Agricultural Water Consumption Based on DWT-WFGM(1,1)-ARMA |
|
Wang Guanzhi1,2, Su Xiaoling1,2, Zhang Te1,2, Jiang Tianliang1,2, Chu Jiangdong1,2
|
|
1.College of Water Resources and Architectural Engineering, Northwest A&2.F University, Yangling
|
| Abstract: |
| 【Objective】Agricultural water consumption prediction is important for regional water resources planning and management. To solve the oscillation of agricultural water consumption and the single result of traditional models, this paper proposed a novel combined forecasting model named DWT-WFGM(1,1)-ARMA which divided the original series into approximate and detailed series by using the Discrete Wavelet Transform.【Method】The Auto regressive Moving Average Model and the Fractional Order Grey Model were applied to predict the detailed series and the approximate series, respectively. Then, the Weighted Markov Chain was adopted for the approximate series error correction, and the predicted results of different component series were added to obtain the predicted results and interval of agricultural water consumption.【Result】Moreover, the proposed model was applied to predict agricultural water consumption in Shaanxi Province and Inner Mongolia Province, and was compared with the traditional grey models including GM(1,1), DWT-GM(1,1)-ARMA and DWT-FGM(1,1)-ARMA. The results showed the mean absolute percentage error of the DWT-FGM(1,1)-ARMA model in Shaanxi Province and Inner Mongolia Province were 1.25% and 1.01% respectively, indicating that the proposed model performed better than other models. And, the proposed model was practical that include compute a reasonable range of fluctuations of agricultural water consumption.【Conclusion】In summary, the model constructed in this paper can effectively improve the accuracy of agricultural water consumption prediction, while the proposed prediction interval can provide a more reliable basis for regional agricultural water consumption prediction. |
| Key words: agricultural water consumption; fractional order grey model; weighted Markov chains; discrete wavelet transform; predicted intervals |
