Stock Market Price Prediction Using Gated Recurrent Unit Method

Abstract

The challenges faced in stock prediction due to intricate dynamics and the lack of interpretability in existing artificial intelligence (AI) models. Traditional models like Long Short-Term Memory (LSTM) and Convolutional Neural Networks (CNN) encounter limitations in capturing long-term trends and adapting to non-stationary data. To address these issues, this research proposes the utilization of Gated Recurrent Unit (GRU) algorithms, which excel in handling sequential data without the complexity of LSTM models. Additionally, the integration of sentiment analysis from news headlines is explored to enhance prediction robustness. By combining GRU models with sentiment analysis, this study aims to provide investors with more accurate predictions and timely notifications, thus improving overall user experience and trust in AI-driven stock market predictions. Through empirical evaluation, the effectiveness of the GRU-based approach in capturing stock market intricacies and enhancing predictability is demonstrated, offering a promising avenue for future research in financial forecasting.