prediction_models

What problem does it solve?

This Skill organizes how the project’s multi-model lottery prediction engine generates recommendations from historical lottery patterns, statistics, and modeled probability distributions.

Core Features & Use Cases

• Multi-model ensemble design (Models A-J): Centralizes different modeling philosophies—statistical similarity, ML classifiers (RF/XGBoost/LightGBM/CatBoost), sequence learning (LSTM), probabilistic state modeling (HMM), extreme-value mean reversion (EVT), Poisson-based omission pressure (J), and heuristic search (GA).
• Model-specific assumptions and targets: Defines what each model tries to learn (e.g., omission pressure, hidden mode transitions, extreme deviation rebound, morphological match fitness) and how it maps to predicted number probabilities.
• Unified execution via the project runner: Documents that all models are dispatched through the common entry point (multi_model.py) using a consistent interface and method selector.

Quick Start

Ask the assistant to “Summarize Models A-J and explain how to run multi_model.py for SSQ using --method A,G,H.”