Comparison between ARIMA and Deep Learning Models for Temperature Forecasting
This is an incremental comparison for weather forecasting applications like agriculture and aviation.
The paper tackled temperature forecasting by comparing ARIMA and a deep learning model, finding that the deep learning model performed better, though no concrete numbers were provided.
Weather forecasting benefits us in various ways from farmers in cultivation and harvesting their crops to airlines to schedule their flights. Weather forecasting is a challenging task due to the chaotic nature of the atmosphere. Therefore lot of research attention has drawn to obtain the benefits and to overcome the challenges of weather forecasting. This paper compares ARIMA (Auto Regressive Integrated Moving Average) model and deep learning models to forecast temperature. The deep learning model consists of one dimensional convolutional layers to extract spatial features and LSTM layers to extract temporal features. Both of these models are applied to hourly temperature data set from Szeged, Hungry. According to the experimental results deep learning model was able to perform better than the traditional ARIMA methodology.