Multi-Confirmation Hybrid Algorithmic Framework

Advanced PhD Research in Intraday Stock Market Prediction using HDQN & Graph Neural Networks.

Research Achievements

0.85
Sharpe Ratio

Significant improvement over baseline (0.50)

90.0%(+/-)5
Classification Accuracy

Achieved via Enhanced GCN Models

0.04%
Forecasting Error

MAPE using Adaptive Recency Weighting

Thesis Abstract

This research integrates Reinforcement Learning agents with Adaptive Forecasting to solve the volatility problem in Indian Financial Markets. The framework functions as a team of three expert advisors, collectively enhancing investment decisions and risk management.