ADAPTIVE DEEP LEARNING MODELS FOR REAL-TIME STOCK MARKET PREDICTION AND AUTOMATED TRADING

Abstract

The rapid evolution of global financial markets has increased the demand for intelligent, data-driven trading systems capable of analyzing complex, high-frequency market dynamics. This study proposes an adaptive deep learning framework for real-time stock market prediction and automated trade execution, integrating multi-modal inputs such as historical price movements, technical indicators, economic signals, and sentiment data. The proposed model utilizes a hybrid architecture combining Long Short-Term Memory (LSTM) networks, Temporal Convolutional Networks (TCN), and reinforcement learning–based policy optimization to enhance predictive accuracy and decisionmaking robustness. A dynamic model-updating mechanism is incorporated to continuously adapt to evolving market behaviors and reduce sensitivity to volatility and noise. Experimental results demonstrate superior performance compared to conventional machine learning and single-model deep learning approaches, achieving improved prediction stability, reduced trading risk, and higher portfolio returns. The framework offers a scalable and intelligent solution for real-time automated trading, contributing significantly to the development of next-generation AI-driven financial systems.