Predicting stock prices of 6 international major corporations: a comparative study of LSTM and FBProphet

Over the years, individuals have been investing in the stock market as a source of passive income. However, such investments are associated with high risk due to the unpredictable behaviour of stock prices. To address this issue, we implemented predictive models to provide investors with stock price predictions to facilitate decision making, reducing loss risk and maximizing profits. A comparative analysis of two models – LSTM (Long Short-Term Memory) and FBProphet – was performed using recent data from six international corporations. This study aims to identify the best-performing model, and performance was evaluated according to each model’s Mean Absolute Error (MAE), Mean Squared Error (MSE) and Root Mean Squared Error (RMSE) metrics. Our objective was to confirm findings in the recent literature, which conclude that LSTM is the best predictive model for stock market prediction. Accordingly, our hypothesis was that LSTM outperforms other algorithms when it comes to stock price sequential data. To confirm our hypothesis, a test statistic was performed, and the results confirmed that LSTM provides superior predictive performance, further validating the literature consensus on the use of LSTM for stock prediction.