Electric Vehicle Stock Price Prediction using LSTM, Bi-LSTM and GRU

Time series analysis, which is used to predict future values from past data, is an indispensable part of decision-making and policy-planning processes. For this reason, this field, which has a wide literature, is widely applied in many areas from economy to engineering, from climate science to health sector. In this study, the application of LSTM (Long Short-Term Memory), BiLSTM (Bidirectional Long Short-Term Memory), and GRU (Gated Recurrent Unit) methods on the stock price dataset of two electric vehicle companies is investigated in order to predict future values. LSTM, Bi-LSTM, and GRU methods are types of RNNs used in time series analysis to identify long-term dependencies in ordered data. The data set is separated into 80% training, and 20% test data. The data is normalized to make accurate predictions and to determine the relationship between past values and future values. MAPE, RMSE, R2, MSE, and MAE performance metrics are used to measure the performance of the data set. The stock predicting models of LSTM, Bi-LSTM, and GRU methods are compared, and the results show that the GRU method significantly outperforms the existing methods.