Powerball Prediction System

A comprehensive machine learning-based system for Powerball lottery number prediction, now refactored into a modular architecture for better maintainability and extensibility.

Issues Fixed

• CSV Header Misalignment: Fixed missing "Powerball" column in powerball.csv headers
• Data Mapping Errors: Corrected incorrect column mapping that was treating 'Day' as 'Year' and misaligning number columns
• Code Organization: Refactored large monolithic file into specialized modules

Modular Architecture

The system is now organized into specialized modules:

Core Modules

1. data_handler.py - Data loading, preparation, and feature engineering

   • PowerballDataHandler class handles CSV loading and data preprocessing
   • Properly maps columns: Month, Day, Year, Num 1-5, Powerball, PowerPlay
   • Engineers comprehensive features for prediction algorithms

2. cache_manager.py - Caching system for model training and features

   • CacheManager class handles model and feature caching
   • Validates cache based on data and model hashes
   • Improves performance by avoiding redundant computations

3. models.py - Neural networks and ensemble model definitions

   • PowerballLSTM - LSTM model for sequence prediction
   • AdvancedPowerballLSTM - Advanced LSTM with attention mechanism
   • EnsembleModelBuilder - Builds ensemble of ML models (RF, XGBoost, etc.)

4. predictors.py - Core prediction algorithms

   • PredictionEngine class with multiple prediction methods
   • ML-based, pattern-based, and hybrid prediction strategies
   • Ensures prediction diversity and uniqueness

5. evaluators.py - Testing and evaluation methods

   • PredictionEvaluator class for comprehensive testing
   • Historical stress testing against past draws
   • Future simulation testing with synthetic draws
   • Performance scoring and statistical analysis

6. visualizers.py - Data visualization and analysis plots

   • PowerballVisualizer class for creating comprehensive charts
   • Frequency heatmaps, temporal analysis, correlation matrices
   • Prediction vs historical comparison plots

7. main_predictor.py - Main orchestrator class

   • AdvancedPowerballPredictor coordinates all components
   • Provides high-level interface for prediction workflows
   • Includes convenience functions and example usage

Quick Start

Simple Usage

from main_predictor import quick_prediction

# Generate 10 predictions quickly
predictions = quick_prediction('powerball.csv', num_predictions=10)

Advanced Usage

from main_predictor import AdvancedPowerballPredictor

# Initialize the system
predictor = AdvancedPowerballPredictor('powerball.csv')

# Generate predictions using different methods
ml_predictions = predictor.generate_predictions(5, method='ml')
pattern_predictions = predictor.generate_predictions(5, method='pattern') 
hybrid_predictions = predictor.generate_predictions(5, method='hybrid')

# Test predictions
test_results = predictor.test_predictions(ml_predictions)

# Create visualizations
predictor.create_visualizations()

# Run comprehensive analysis
results = predictor.run_comprehensive_analysis()

Individual Component Usage

# Use components individually
from data_handler import PowerballDataHandler
from predictors import PredictionEngine
from evaluators import PredictionEvaluator

data_handler = PowerballDataHandler('powerball.csv')
prediction_engine = PredictionEngine(data_handler)
evaluator = PredictionEvaluator(data_handler)

# Generate features and predictions
features = data_handler.engineer_features()
predictions = prediction_engine.generate_pattern_based_predictions(features)
results = evaluator.stress_test(predictions)

Data Format

The system expects a CSV file with the following headers:

Month,Day,Year,Num 1,Num 2,Num 3,Num 4,Num 5,Powerball,PowerPlay (if applicable)

Example data row:

10,3,2015,33,6,46,44,26,4,2

Where:

• Month, Day, Year: Draw date
• Num 1-5: White ball numbers (1-69)
• Powerball: Red powerball number (1-26)
• PowerPlay: Multiplier (optional)

Files

Original (deprecated)

• powerballPredictions.py - Original monolithic implementation

New Modular System

• data_handler.py - Data loading and preprocessing
• cache_manager.py - Caching system
• models.py - ML models and neural networks
• predictors.py - Prediction algorithms
• evaluators.py - Testing and evaluation
• visualizers.py - Data visualization
• main_predictor.py - Main orchestrator
• example_usage.py - Usage examples

Data

• powerball.csv - Historical Powerball draws (headers now corrected)

Generated

• model_cache/ - Cached models and features
• stress_test_results/ - Test results and performance data
• visualizations/ - Generated charts and plots

Features

• Multiple Prediction Methods: ML-based, pattern-based, and hybrid approaches
• Comprehensive Testing: Historical and future simulation stress testing
• Advanced Analytics: Feature engineering, statistical analysis, correlation studies
• Visualization: Interactive plots for data exploration and prediction analysis
• Caching: Intelligent caching for improved performance
• Modular Design: Easy to extend and maintain

Performance

The system has been optimized for:

• Fast prediction generation (< 5 seconds for 10 predictions)
• Efficient feature engineering with caching
• Comprehensive testing against historical data
• Memory-efficient model training and inference

Requirements

• Python 3.7+
• pandas, numpy, scikit-learn
• xgboost, lightgbm
• pytorch (for neural networks)
• plotly (for visualizations)
• statsmodels, prophet (for time series analysis)

Note

This system is for educational and entertainment purposes. Lottery numbers are random and no prediction system can guarantee winning numbers.