Trading Strategy Intern
• Developed python functions to generate and analyse 10+ new features for Open to Close trading strategy.• Conducted over 12+ experiments to optimize O2C model performance such as model variance analysis, addingNIFTY sectoral indices and commodities to base model, feature selection using embedded methods.• Improved short-class label prediction accuracy by 20% for C2O strategy through techniques like SMOTE, BalancedRandom Forest, and hyperparameter tuning.• Performed correlation analysis, enabling identification of 2 key features linked to model’s predicted returns.