Stock-Forecasting-with-Machine-Learning

Deployed a Gradient Boosting model that generated a 7x lift in ROI vs baseline — proving machine learning isn't just theoretical, it's profitable.

Model Benchmarking

This project tested every mainstream model for forecasting market Direction]:

• Logistic Regression
• Random Forests
• Decision Trees
• Neural Networks (MLP/Keras)
• Gradient Boosting

Result: Gradient Boosting delivered 7x higher ROI on the holdout dataset vs average baseline.

Why Gradient Boosting Won:

• Handles feature interactions automatically
• Deals with non-linearity & noise better
• Outperformed even MLP in limited-data setting