基于ARIMA模型和LSTM模型对南极地表月平均温度的预测  

Prediction of antarctic monthly mean surface temperature based on ARIMA model and LSTM model

作　　者：相旭东 胡学平 余晓美 邓新 XIANG Xudong;HU Xueping;YU Xiaomei;DENG Xin(School of Mathematics and Finance,Chuzhou College,Chuzhou 239000,China;School of Mathematics and Physics,Anqing Normal University,Anqing 246133,China)

机构地区：[1]滁州学院数学与金融学院,安徽滁州239000 [2]安庆师范大学数理学院,安徽安庆246133

出　　处：《高师理科学刊》2021年第6期33-37,共5页Journal of Science of Teachers＇College and University

基　　金：安徽省自然科学研究项目(KJ2018A0424);滁州学院科研启动基金资助项目(2018qd01);安徽省自然科学基金青年项目(1908085QA01);应用统计学专业卓越工程师培养创新项目(2020zyrc089);校级一流本科专业项目(2019aqnuylzy01);校级一流教材建设项目(2019 aqnuylkc13)。

摘　　要：全球气候变暖对自然环境和人类活动造成了极为恶劣的影响,研究南极地表气温的变化规律能够帮助人类更好地应对这一问题.以Marion站点1968—2020年的地表月平均温度数据作为研究对象,分别采用传统的ARIMA模型和神经网络中的LSTM模型进行分析预测.对比RMSE和MAE发现,LSTM模型的拟合效果更好,因此对于掌握气温未来发展趋势具有重要价值.Global warming has a very bad impact on the natural environment and human activities,studying the variation of antarctic surface temperature can help people deal with this problem better.The monthly mean surface temperature data of Marion station from 1968 to 2020 are taken as the research object.The traditional ARIMA model and the LSTM model of neural network are used for analysis and prediction.By comparing RMSE and MAE,it is found that the fitting effect of LSTM model is better,which has important value for mastering the future development trend of temperature.

关 键 词：地表气温 ARIMA模型 LSTM模型 南极 

分 类 号：O212∶S161.23[理学—概率论与数理统计]