Comparative analysis of VARMA method and LSTM in predicting stock price

Nowadays, financial markets are affected by various social and political events. This makes forecasting even more important. Forecasting financial markets help stockholders to sell or buy the stock, when necessary, which leads to increased profits. Since there are many forecasting methods, this paper applies two methods, Vector Autoregressive Moving Average (VARMA) and Long Short-Term Memory (LSTM) for forecasting. Forecasting was done on the stock price data of two Apple and Microsoft companies, which include four features: Open, High, Close, and Low. By comparing these two prediction models using Mean Square Error (MSE), Root Mean Square Error (RMSE), and Mean Absolute Error (MAE) criteria, it is concluded that LSTM method is more suitable for stock price prediction than VARMA.

Keywords:

Stock price , Vector autoregressive moving average , Recurrent neural network , Long short-term memory.