Automated AI trading system

Abstract

This Work Project is an empirical investigation and a prototype development of an automated AI trading system. The framework system is a fully automated pipeline of data processing, data analytics, and signals interpretation that ends with an action to buy, hold, or sell an asset. The data analytics segment represents a deployment of state-of-the-art AI models developed to predict future cryptocurrency prices while accounting for risk and order management. The framework proposed predicted the next 15-minuteclose price of Bitcoin achieving an RMSE value of 167during the period of 11th-15thDecember. After accounting for fees and commissions, the system would have yielded a return of (1.11%) with a Sharpe Ratio of (0.04)against a B&H strategy with (3.92%)and (0.09)Sharpe.