Comparative Analysis of LSTM, Transformer-Based LSTM and Transformer-Based GRU for Stock Price Prediction: An Implementation in Vietnam Market
摘要
The volatility of stock prices in emerging markets as Vietnam presents challenges in forecasting precision prices. In recent years, many deep learning models that can capture time-series data with high accuracy have been developed. This study proposed a pipeline to forecast stock prices, including implementation of state-of-the-art Transformer architectures, evaluating the efficiency, imposing a comparison of LSTM, Transformer-based LSTM and Transformer-based GRU applied in Vietnam’s Equity market. Data was collected from five leading stocks in the period of five years (2019–2024). The results showed that the Transformer-based LSTM and Transformer-GRU exhibited exceptional results regarding some stock codes. The study emphasized the role of technologies on modern finance especially in Vietnam stock market, which is forecasted to experience robust growth in 2025.