Petrobras PETR3 stock price forecast: an application of CNN-LSTM neural networks

Abstract

Abstract

This work proposes using the CNN-LSTM network to predict the opening price of Petrobras PETR3 stock. The database offers a daily series of PETR3 stock prices between January/2012 and December/2022, totaling 2713 observations. Multivariate prediction models based on the CNN-LSTM (Convolutional Neural Network - Long Short Term Memory), LSTM (Long Short Term Memory), GRU (Gated Recurrent Unit), and CNN (Convolutional Neural Network) networks were implemented in the Python language. Results obtained from the four models were compared using the metrics RSME (Root Mean Squared Error), MAPE (Mean Absolute Percent Error), and MAE (Mean Absolute Error). It was verified for a 30-day horizon that the CNN-LSTM model presented a better performance.