i-Trader: Intelligent Trading Bot

i-Trader: Intelligent Trading Bot

Sarvesh Natkar, Om Melkeri, Dr. Pratyush Urade

Department of Artificial Intelligence and Data Science,

New Horizon Institute of Technology & Management Thane, India

Abstract—The volatility and unpredictability of financial mar- kets pose a significant challenge for traditional trading strategies, which often fail to adapt to rapid market shifts. This paper presents a Reinforcement Learning (RL)-based trading bot utilizing Proximal Policy Optimization (PPO) — a powerful Deep Reinforcement Learning algorithm — to optimize stock trading decisions. The bot integrates with the Alpaca API for real-time market data and paper trading execution, ensuring practical applicability. It evaluates stocks using technical indicators and dynamically learns from market patterns to maximize profitability. Backtesting on tech sector stocks (Apple, Microsoft, Nvidia) demonstrated an 18 percent return, outperforming standard buy- and-hold strategies. The results validate the bot’s ability to adapt, handle market fluctuations, and execute trades efficiently. Future enhancements, including multi-stock portfolio management and sentiment analysis integration, are proposed to further improve performance and market adaptability.

Index Terms—Deep Reinforcement Learning (DRL), Proxi- mal Policy Optimization (PPO),Algorithmic Trading,Financial Market Prediction, Market Trend Analysis, Automated Portfolio Management.