基于CNN-Attention-BP的降水发生预测研究
作者:
中图分类号:

P457.6

基金项目:

国家重点研发计划(2018YFC1507905);国家自然科学基金(42075068,41975087);2020年江苏高校"大学素质教育与数字化课程建设"专项课题(2020JDKT032);南京信息工程大学2019年教改研究课题(共建共享的概率论与数理统计"金课"的探索与实践);南京信息工程大学数统学院本科专业建设项目


Application of CNN-Attention-BP to precipitation forecast
Author:
  • 摘要
  • | |
  • 访问统计
  • |
  • 参考文献
  • |
  • 相似文献
  • | | |
  • 文章评论
    摘要:

    在综合分析降水统计预测模型特点的基础上,提出一种基于Attention机制、卷积神经网络(CNN)和BP神经网络的CNN-Attention-BP组合模型,并对1961—2020年不同气候类型的长春站、白城站、延吉站夏季降水进行实证分析.首先,运用卷积神经网络对6—8月20—次日20时降水量、平均气压、平均风速、平均气温和平均相对湿度进行特征学习,利用Attention机制来确定气象影响因素对降水预测的权重;然后,使用BP神经网络进行降水发生预测,选用准确率、交叉熵损失函数和F1-score来综合评价CNN-Attention-BP组合模型的性能.最后,将单一的支持向量机、多层感知机和卷积神经网络模型与组合模型进行比较分析.结果表明,CNN-Attention-BP组合模型具有自主学习和关注更重要信息的特征,能够有效提高吉林省夏季降水发生模型的预测能力,在样本越均衡、降水频率越接近于0.5的站点,预测精度越高,准确率最高可达88.4%.CNN-Attention-BP组合模型的准确率相较于其他单一模型最高可以提高近17个百分点.

    Abstract:

    Here,a CNN-Attention-BP model is proposed for precipitation forecast based on analysis of the characteristics of precipitation statistical prediction models,then empirical analyses are made on the summer rainfall in Changchun,Baicheng and Yanji stations of Jilin province for the period of 1961-2020.First,a Convolution Neural Network (CNN) is used to study the characteristics of precipitation,air pressure,wind speed,air temperature and relative humidity.Second,the attention mechanism is used to determine the weight of meteorological factors for precipitation forecast.Then a BP neural network is applied to predict the precipitation probability.And the performance of the proposed CNN-Attention-BP model is evaluated by accuracy,cross-entropy loss function and F1-score,which is compared with that of the support vector machine,multi-layer perceptron and convolution neural network model.The results show that the CNN-Attention-BP model is characterized by autonomous learning and paying more attention to significant information,as well as the improved forecasting performance in summer precipitation occurrence for Jilin province.Meanwhile,the proposed model would perform better with more balanced sample and precipitation frequency closer to 0.5,when accuracy would reach up to 88.4%.Compared with single models,the CNN-Attention-BP improves the forecast accuracy by 17 percentage points.

    参考文献
    [1] 王晓东,董炜华,褚丽娟,等.吉林西部旱涝灾害与降水对应关系的分析[J].灾害学,2015,30(4):53-58 WANG Xiaodong,DONG Weihua,CHU Lijuan,et al.Relationship analysis of drought-flood disasters and precipitation in western Jilin province[J].Journal of Catastrophology,2015,30(4):53-58
    [2] Ramana R V,Krishna B,Kumar S R,et al.Monthly rainfall prediction using wavelet neural network analysis[J].Water Resources Management,2013,27(10):3697-3711
    [3] Radhakrishnan P,Dinesh S.An alternative approach to characterize time series data:case study on Malaysian rainfall data[J].Chaos,Solitons&Fractals,2006,27(2):511-518
    [4] Barua S,Muttil N,Ng A W M,et al.Rainfall trend and its implications for water resource management within the Yarra River catchment,Australia[J].Hydrological Processes,2013,27(12):1727-1738
    [5] 张芳华,曹勇,徐珺,等.Logistic判别模型在强降水预报中的应用[J].气象,2016,42(4):398-405 ZHANG Fanghua,CAO Yong,XU Jun,et al.Application of the logistic discriminant model in heavy rain forecasting[J].Meteorological Monthly,2016,42(4):398-405
    [6] 张佳华,姚宜斌,曹娜.基于决策树对有无降水进行预测[J].测绘地理信息,2017,42(5):107-109 ZHANG Jiahua,YAO Yibin,CAO Na.Prediction of whether precipitation based on decision tree[J].Journal of Geomatics,2017,42(5):107-109
    [7] Ortiz-Garcia E G,Salcedo-Sanz S,Casanova-Mateo C.Accurate precipitation prediction with support vector classifiers:a study including novel predictive variables and observational data[J].Atmospheric Research,2014,139:128-136
    [8] 黎玥君,郭品文.基于BP神经网络的浙北夏季降尺度降水预报方法的应用[J].大气科学学报,2017,40(3):425-432 LI Yuejun,GUO Pinwen.Application of downscaling forecast for the north of Zhejiang precipitation in summer based on the BP neural network model[J].Transactions of Atmospheric Sciences,2017,40(3):425-432
    [9] 张鹏程,贾旸旸.一种基于多层感知器的动态区域联合短时降水预报方法[J].计算机应用与软件,2018,35(11):153-158,183 ZHANG Pengcheng,JIA Yangyang.A dynamic regional combined short-term precipitation forecasting approach based on multilayer perceptron[J].Computer Applications and Software,2018,35(11):153-158,183
    [10] 李增.东北地区干旱特征及预测模型研究[D].沈阳:沈阳农业大学,2021 LI Zeng.Study on drought characteristics and prediction model in Northeast China[J].Shenyang:Shenyang Agricultural University,2021
    [11] Pan B X,Hsu K,AghaKouchak A,et al.Improving precipitation estimation using convolutional neural network[J].Water Resources Research,2019,55(3):2301-2321
    [12] 刘嘉琛,秦小麟,朱润泽.基于LSTM-Attention的RFID移动对象位置预测[J].计算机科学,2021,48(3):188-195 LIU Jiachen,QIN Xiaolin,ZHU Runze.Prediction of RFID mobile object location based on LSTM-Attention[J].Computer Science,2021,48(3):188-195
    [13] Junninen H,Niska H,Tuppurainen K,et al.Methods for imputation of missing values in air quality data sets[J].Atmospheric Environment,2004,38(18):2895-2907
    [14] Rawat W,Wang Z H.Deep convolutional neural networks for image classification:a comprehensive review[J].Neural Computation,2017,29(9):2352-2449
    [15] Zang H X,Liu L,Sun L,et al.Short-term global horizontal irradiance forecasting based on a hybrid CNN-LSTM model with spatiotemporal correlations[J].Renewable Energy,2020,160:26-41
    [16] Liang Z Y,Du J P,Li C Y.Abstractive social media text summarization using selective reinforced Seq2Seq attention model[J].Neurocomputing,2020,410:432-440
    [17] 谢浩.基于BP神经网络及其优化算法的汽车车速预测[D].重庆:重庆大学,2014 XIE Hao.Prediction of driving condition for plug-in hybrid electric vehicles[D].Chongqing:Chongqing University,2014
    [18] 李静,徐路路.基于机器学习算法的研究热点趋势预测模型对比与分析:BP神经网络、支持向量机与LSTM模型[J].现代情报,2019,39(4):23-33 LI Jing,XU Lulu.Comparison and analysis of research trend prediction models based on machine learning algorithm:BP neural network,support vector machine and LSTM model[J].Modern Information,2019,39(4):23-33
    [19] 张瀚超.基于神经网络组合模型的风电场短期功率预测方法研究[D].株洲:湖南工业大学,2019 ZHANG Hanchao.Research on short-term prediction of wind power based on optimized neural network[D].Zhuzhou,China:Hunan University of Technology,2019
    [20] Chang Z H,Zhang Y,Chen W B.Electricity price prediction based on hybrid model of Adam optimized LSTM neural network and wavelet transform[J].Energy,2019,187:115804
    相似文献
    引证文献
    网友评论
    网友评论
    分享到微博
    发 布
引用本文

吴香华,华亚婕,官元红,王巍巍,刘端阳.基于CNN-Attention-BP的降水发生预测研究[J].南京信息工程大学学报(自然科学版),2022,14(2):148-155
WU Xianghua, HUA Yajie, GUAN Yuanhong, WANG Weiwei, LIU Duanyang. Application of CNN-Attention-BP to precipitation forecast[J]. Journal of Nanjing University of Information Science & Technology, 2022,14(2):148-155

复制
分享
文章指标
  • 点击次数:521
  • 下载次数: 1487
  • HTML阅读次数: 280
  • 引用次数: 0
历史
  • 收稿日期:2021-05-31
  • 在线发布日期: 2022-04-27

地址:江苏省南京市宁六路219号    邮编:210044

联系电话:025-58731025    E-mail:nxdxb@nuist.edu.cn

南京信息工程大学学报 ® 2025 版权所有  技术支持:北京勤云科技发展有限公司