Bayesian-Driven Investment Strategy: Integrating LSTM Forecasting, Market State Estimation, and Portfolio Optimization

H. Sun

We propose an automated investment system that integrates Bayesian Inference, Reinforcement Learning (RL), and Long Short-Term Memory (LSTM) for dynamic stock trading. The framework continuously forecasts stock returns using LSTM, while Bayesian Inference refines predictive distributions based on new market data, capturing uncertainty and adjusting trading signals accordingly.

A Bayesian Hedging Strategy translates investment signals into adaptive trading actions, balancing expected returns with market risk. RL further optimizes the strategy by learning an optimal trading policy through historical market data, refining decisions via trial and error. The system automates the entire pipeline, including real-time data collection, forecasting, decision-making, and trade execution via broker APIs.

This work advances autonomous trading systems, leveraging Bayesian decision-making and reinforcement learning for more adaptive, uncertainty-aware investment strategies.

Keywords: Bayesian Inference, Reinforcement Learning, LSTM, Stock Prediction

Interdisciplinary applications of Bayesian methods
June 10, 2025  3:30 PM
Sala de prensa (MR 13)

J. M. Camacho Rodríguez, R. Naveiro, D. Rios Insua

P. García Arce, R. Naveiro Flores, D. Ríos Insua

R. Naveiro, M. Carreau, W. N. Caballero