Global Temperature Prediction based on Time Series and Global Climate

Abstract

Global warming is one of the topics of the highest public concern in recent years. This paper mainly analyzes and forecasts the global temperature. This paper have established three models to describe the past and predict the future global temperature level. The first model is the integrated model of ARIMA and BP neural network, which can flexibly mine the linear and nonlinear relationships behind the data. The second model is the Full Convolution Network (FCN) model, which can predict end-to-end time series. The third model is the SVR-QM model based on climate indicators. The three models are established at different angles, and all have good fitting effects.