Multi-Agent Stock Prediction Systems: Machine Learning Models, Simulations, and Real-Time Trading Strategies

TL;DR

The paper tackles the challenge of accurate stock price forecasting by systematically comparing multiple neural architectures, including LSTM, GRU, and Transformer-based attention models, on stock time-series data. It conducts a thorough experimental study using Tesla stock data, evaluating single-layer configurations with various architectural enhancements such as bidirectionality and Seq2Seq frameworks. The key finding is that attention-based models achieve the highest accuracy (up to approximately 95%), with bidirectional LSTM and Seq2Seq variants also performing strongly, while GRU variants generally underperform relative to LSTMs. The results offer practical guidance for deploying AI-driven trading systems and motivate future work on hybrid architectures and real-time, multi-asset deployment.

Abstract

This paper presents a comprehensive study on stock price prediction, leveragingadvanced machine learning (ML) and deep learning (DL) techniques to improve financial forecasting accuracy. The research evaluates the performance of various recurrent neural network (RNN) architectures, including Long Short-Term Memory (LSTM) networks, Gated Recurrent Units (GRU), and attention-based models. These models are assessed for their ability to capture complex temporal dependencies inherent in stock market data. Our findings show that attention-based models outperform other architectures, achieving the highest accuracy by capturing both short and long-term dependencies. This study contributes valuable insights into AI-driven financial forecasting, offering practical guidance for developing more accurate and efficient trading systems.

Figures (14)

• Figure 1: LSTM cell gfglstm
• Figure 2: GRU cell huang2019convolutional
• Figure 3: Transformer architecture
• Figure 4: LSTM
• Figure 5: LSTM-2 Path