ML4T Trading Bot

Abstract

Algorithmic trading is a useful test bed for machine-learning systems because it forces decisions under uncertainty, feedback loops, and imperfect signals.

This OMSCS project explored trading strategies built with reinforcement-learning and decision-tree ideas, with emphasis on evaluation, comparison, and failure-case reasoning rather than raw returns alone.

What I Built

• The project compared multiple trading-strategy paths instead of relying on a single heuristic.
• Its strongest signal is disciplined evaluation and reasoning under uncertainty rather than headline model output.

Impact

• Shows applied machine-learning thinking in a finance context with clear ties to experimentation and strategy design.
• Works as a concise OMSCS case study for decision-oriented AI work.