Forecasting Models
ARIMA Model
Statistical time-series model that captures trends and seasonality. Best for stable, trending markets.
Use when: Clear trend with low volatility
LSTM Model
Deep learning neural network that learns complex patterns from historical data. Powerful but can overfit.
Use when: Complex patterns with sufficient data
Prophet Model
Facebook's decomposition model that handles seasonality and holidays well. Good for long-term forecasts.
Use when: Long-term predictions needed
Ensemble Model
Combines all three models for balanced, robust predictions. Reduces individual model bias.
Use when: Uncertainty is high or market regime unclear