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