Bruno Giorgio | Searcharxiv

arXiv2025

Phase Transitions in Financial Markets Using the Ising Model: A Statistical Mechanics Perspective

This dissertation investigates the ability of the Ising model to replicate statistical characteristics, or stylized facts, commonly observed in financial assets. The study specifically examines in the S&P500 index the following features: volatility clustering, negative skewness, heavy tails, the absence of autocorrelation in returns, and the presence of autocorrelation in absolute returns. A significant portion of the dissertation is dedicated to Ising model-based simulations. Due to the lack of an analytical or deterministic solution, the Monte Carlo method was employed to explore the model's statistical properties. The results demonstrate that the Ising model is capable of replicating the majority of the statistical features analyzed.

q-fin.ST↗

arXiv2025

Dueling Deep Reinforcement Learning for Financial Time Series

Reinforcement learning (RL) has emerged as a powerful paradigm for solving decision-making problems in dynamic environments. In this research, we explore the application of Double DQN (DDQN) and Dueling Network Architectures, to financial trading tasks using historical SP500 index data. Our focus is training agents capable of optimizing trading strategies while accounting for practical constraints such as transaction costs. The study evaluates the model performance across scenarios with and without commissions, highlighting the impact of cost-sensitive environments on reward dynamics. Despite computational limitations and the inherent complexity of financial time series data, the agent successfully learned meaningful trading policies. The findings confirm that RL agents, even when trained on limited datasets, can outperform random strategies by leveraging advanced architectures such as DDQN and Dueling Networks. However, significant challenges persist, particularly with a sub-optimal policy due to the complexity of data source.

cs.LG↗