ml-alpha

What problem does it solve?

This Skill addresses the complex challenge of generating and validating machine-learning based alpha signals in liquid markets, ensuring robustness through leakage-safe training and execution-aware evaluation.

Core Features & Use Cases

• Supervised and Ensemble Models: Develops alpha models using both supervised learning and ensemble techniques.
• Leakage-Safe Training: Implements safeguards to prevent data leakage during the training process.
• Execution-Aware Evaluation: Evaluates model performance considering realistic trading costs and execution assumptions.
• Use Case: A quantitative analyst needs to develop a new predictive model for stock price movements. They can use this Skill to engineer features, train the model with robust validation, and evaluate its potential profitability after accounting for transaction costs.

Quick Start

Run the ml alpha diagnostics script on the input CSV file to generate a diagnostics JSON artifact.