中文摘要：

运用EVIEWS软件，对铜陵市48年来的月平均气温时间序列进行统计分析，并对该动态数据进行建模和预测。采用差分方法对样本数据进行预处理，然后定阶，并进行参数估计，建立季节ARIMA模型对铜陵市气温数据进行预报。预报结果显示，季节ARIMA模型的平均绝对误差值为0．875。将ARIMA模型预报结果与径向基（radial basis function，RBF）神经网络模型的预报值比较可知，其预报结果优于RBF神经网络的预测结果。

英文摘要：

In this article, we analyzed the time series data of month mean temperature in Tongling city us- ing EVIEWS software, and got modeling prediction according to the dynamical data. We preprocessed the sample data using difference method, and then made sure the model order and estimated the parameter values for estab- lishing the season autoregressive integrated moving average （ARIMA） model to fit the time series. The prediction results showed that the average absolute error of the season ARIMA model is 0.875. Comparing the result of ARIMA model and RBF （radial basis function） neural network, the season ARIMA model is better than radial basis function （RBF） neural network.