کنفرانس مهندسی برق ایران

صفحه اصلی / سی و سومین کنفرانس بین المللی مهندسی برق

An Integrated Technical Analysis and Machine Learning Trading Model for Noisy and Volatile Financial Markets

نویسندگان :

Arvin Esfandiari¹ Ali Doustmohammadi²

1- دانشگاه صنعتی امیرکبیر(پلی تکنیک تهران) 2- دانشگاه صنعتی امیرکبیر(پلی تکنیک تهران)

کلمات کلیدی :

Financial Markets،xLSTM،RL،MACD Indicator Optimization

چکیده :

Financial markets are notably challenging to predict due to their complex dynamics and low signal-to-noise ratio (SNR). While machine learning methods have been widely explored for financial forecasting, many fail to deliver consistent profitability. This study introduces a novel trading framework that integrates extended long-short-term memory (xLSTM) with a wavelet denoising for improved prediction accuracy. Additionally, a Genetic Algorithm (GA) combined with an Expectation Maximization-Kalman Filter (EM-KF) is employed to optimize the coefficients of the Moving Average Convergence Divergence (MACD) indicator. The predicted outputs and optimized MACD histogram are fed into a Reinforcement Learning (RL) model, which utilizes a Q-table for decision-making. Backtesting on hourly historical trading data for the specific cryptocurrencies demonstrates the robustness of the proposed trading model in terms of profitability.