| 引用本文: | 李勇 ,陈栏灵 ,李禹锋.基于 GPR 模型的气象因素对经济高质量发展的预测———以重庆市为例(J/M/D/N,J:杂志，M：书，D：论文，N：报纸).期刊名称,2024，41（5）：110-118 |
| CHEN X. Adap tive slidingmode contr ol for discrete2ti me multi2inputmulti2 out put systems[ J ]. Aut omatica, 2006, 42(6): 4272-435 |
|
| 摘要: |
| 目的 针对经济社会与气象变化之间的联系越来越密切的现象,以及气象数据、经济高质量发展数据的复杂
特征和传统模型的预测精度不足问题,提出从气象和经济高质量发展关联的视角出发,以统计学方法进行气象因
素对经济高质量发展的预测。 方法 鉴于高斯过程回归模型对于高度非线性回归问题有很强的适应性,同时还能自
适应获取最优超参数,并给出具有概率意义的预测结果,故将高斯过程回归模型引入气象对经济高质量发展的预
测,采用 7 种不同核函数,并分别训练出最优超参数,通过均方误差比较择出预测效果最好的模型核函数及相应参
数。 结果 对重庆市气象与经济高质量发展历史观测数据构建高斯过程回归( GPR) 模型,进行 GPR 建模,并进行
预测误差分析,得到的结果表明:选用参数为 8. 091 的常值核与缩放参数为 9. 454 5 的 RBF 核组合而成的混合核
作为最佳核函数的 GPR 模型,相较于 K 邻近回归模型、支持向量回归模型,误差更低,GPR 模型预测点的 y 值绝对
误差最大为 0. 548,最小为 0. 094,较为准确;模型真实值与预测值对比显示拟合效果较为良好。 结论 GPR 模型运
用于气象因素对经济高质量发展的预测分析具有优良性,并针对气象与经济高质量发展指数的关系特征,提出了
加强气象预报、提高利用效率和精准化预测的有效建议。 |
| 关键词: 气象因子 经济高质量发展 高斯过程回归( GPR) |
| DOI： |
| 分类号: |
| 基金项目: |
|
| Prediction of Meteorological Factors on High-quality Economic Development Based on GPR Model TakingChongqing Municipality as an Example |
|
LI Yong, CHEN Lanling, LI Yufeng
|
|
1. School of Statistics Chengdu University of Information Engineering Chengdu 610225 China
2. School of Mathematics and Statistics Guangxi Normal University Guangxi Guilin 541004 China
|
| Abstract: |
| Objective Aiming at the phenomenon of the increasingly close connection between economy and society and
meteorological changes the complex characteristics of meteorological data and high-quality economic development data
and the problem of insufficient prediction accuracy of the traditional model this study proposed to carry out the prediction
of meteorological factors on the high-quality development of the economy from the perspective of the correlation between
meteorology and high-quality development of the economy by statistical methods. Methods Gaussian process regression model has strong adaptability to highly nonlinear regression problems and it can also adaptively obtain the optimal
hyperparameters and give probabilistic prediction results. Therefore the Gaussian process regression model was introduced
into the prediction of meteorological factors on high-quality economic development. Seven different kernel functions were
used the optimal hyperparameters were trained and the best model kernel function and corresponding parameters were
selected by comparing the mean square errors. Results A Gaussian process regression GPR model was constructed
based on the historical observation data of meteorology and high-quality economic development in Chongqing. The results
obtained from the analysis of prediction errors showed that compared with K-proximity regression model and support vector
regression model the GPR model which used the hybrid kernel formed by the combination of constant kernel with a
parameter of 8. 091 and RBF kernel with scaling parameter of 9. 454 5 as the best kernel function had a lower error. The
absolute error of the y-value predicted by the GPR model was 0. 548 at the maximum and 0. 094 at the minimum which
was more accurate. The comparison between the real values of the model and the predicted values showed a relatively good
fitting effect. Conclusion The GPR model is excellent for the forecast analysis of meteorological factors on high-quality
economic development. For the characteristics of the relationship between meteorology and high-quality economic
development index effective suggestions including strengthening the meteorological forecast improving the utilization
efficiency and making accurate predictions are put forward. |
| Key words: meteorological factor high-quality economic development Gaussian process regression GPR |
