Oil Price Forecast Using Deep Learning and ARIMA

Multiple factors affect crude oil price including economic cycle, international relations, geopolitics etc. Forecasting oil price is a challenging but rewarding task. In this study, we examine the time-series and non-linear feature of oil price and use both time-series model and neural network model to make the prediction. We purpose data preprocessing methods to fit the data into models. The results show that a convolution neural networks yields best performance with relatively simple structure and lesser training time.