基于灰色预测与BP神经网络的全球温度预测研究  

Global Temperature Prediction Based on Grey Prediction and BP Neural Network

作　　者：朱圣耀 陈劲杰[1] ZHU Sheng-yao;CHEN Jin-jie(School of Mechanical Engineering,University of Shanghai for Science and Technology,Shanghai 200093,China)

机构地区：[1]上海理工大学机械学院,上海200093

出　　处：《软件导刊》2020年第2期105-109,共5页Software Guide

摘　　要：为深入了解全球变暖缘由及影响,探究变暖是否停滞,考虑地球吸热、散热及海洋温度变化等因素,构建一种全球温度预测模型,预测未来25年温度变化。采用主成分分析法找出贡献度较大的3个主成分,再用3组灰色预测模型预测海洋平均温度、二氧化碳排放量、太阳长波辐射等8个变量,并进行光滑比、级比和残差检验。结果表明,预测符合前29年(1990-2018年)时间序列图规律,用历史数据训练BP神经网络,然后把8个变量的预测值代入神经网络,拟合优度为0.92272,精度非常高,可以看出全球平均温度距平序列越来越大,说明温度正逐渐升高,而不是停滞。In order to better understand the causes and impacts of global warming,by considering the endothermic,radiating and ocean temperature changes of the earth,a global temperature prediction model is established to predict the temperature change in the next 25 years.The principal component analysis method is used to find the three principal components with large contribution,and then three sets of gray prediction models are used to predict the 8 variables including the ocean average temperature,carbon dioxide emissions,solar long wave radiation,etc,and smoothness ratio,step ratio and residual error tests were made,which proved that the predictions were consistent with the rules of the time series chart of the previous 29 years.The historical data was used to train the BP neural network,and then the 8 predicted values of the variable were substituted into the neural network,and the goodness of fit was 0.92272.The accuracy is very high.It can be seen that the global average temperature anomaly series is getting larger and larger,indicating that the temperature is not stagnant and gradually increasing.

关 键 词：全球变暖 主成分分析 灰色预测 BP神经网络 

分 类 号：TP301[自动化与计算机技术—计算机系统结构]