PREDICTIVE ANALYSIS OF STOCK MARKET PRICES: A COMPARATIVE STUDY OF LSTM AND ARIMAX IN FORECASTING CLOSING PRICES OF EQUITIES

Abstract

The prediction of stock prices poses a substantial challenge in the dynamic landscape of financial markets, capturing the interest of both researchers and investors. This research paper proposes the comparative analysis of two prominent models—Long Short-Term Memory (LSTM) and Auto Regressive Integrated Moving Average with eXogenous variables (ARIMAX)—to uncover their predictive capabilities. Utilizing datasets from technical giants Apple and Samsung, the research employs a systematic methodology, emphasizing the application of these models in predicting stock prices. With preprocessing of the datasets, model training and validation using historical stock prices the future predictions are made. The comparative analysis incorporates plots to unveil unique patterns and dependencies within the data. The outcomes are presented through visualizations and quantitative metrics such as MAE, MSE, and RMSE, offering an understanding of the models' performance on both the datasets.