Forecasting Models¶
Practical how-tos for point forecasters, interval forecasters, class-probability forecasters, and ensembles. Covers gradient boosting integration, quantile regression, conformal methods, and combining multiple models.
-
How to Forecast with CatBoost
Plug CatBoostRegressor into PointReductionForecaster as a drop-in sklearn estimator, compare gradient-boosted versus Ridge linear baseline, and demonstrate the direct reduction strategy with tree-based models.
-
How to Forecast Intervals with CatBoost Multiquantile
Use IntervalReductionForecaster with CatBoost's native multiquantile objective for simultaneous lower and upper bound estimation.
-
How to Forecast Class Probabilities
Use ClassProbaReductionForecaster to produce calibrated probability forecasts and evaluate them with Brier score, log loss, and accuracy.
-
How to Choose a Decomposition Strategy
Build 2- and 3-component DecompositionPipeline forecasters chaining trend, seasonality, and residual models with target pre-transformation.
-
How to Use Distance-Based Similarity for Intervals
Adaptive prediction intervals via similarity-weighted conformal prediction using DistanceSimilarity with configurable distance metrics and bandwidths.
-
How to Build a Lag-Feature Forecaster
Chain feature and target forecasters with ForecastedFeatureForecaster when exogenous variables are unknown at prediction time and must be forecasted.
-
How to Use Lagged Forecasts as Features
Compare ForecastedFeatureForecaster strategies (actual, predicted, rewind) and split ratio tuning for chaining feature and target forecasters.
-
How to Build Interval Forecasts with Reduction
Wrap any quantile-capable sklearn estimator with IntervalReductionForecaster to produce calibrated prediction intervals across multiple horizons.
-
How to Produce Multi-Vintage Predictions
Generate multiple predictions from different weather forecast vintages without refitting, using the X_forecast predict-time override.
-
How to Apply Time-Weighted Training
Use time_weight and sample_weight_alignment to emphasise recent or seasonal training samples in PointReductionForecaster, with visualisation of weight curves and alignment strategy comparison.
-
How to Combine Classification Forecasters
Build classification ensembles with VotingClassProbaForecaster using soft and hard voting strategies.
-
How to Combine Interval Forecasters
Build interval ensembles with VotingIntervalForecaster using envelope, mean, and median aggregation strategies.
-
How to Combine Forecasters with VotingPointForecaster
Build point ensembles with VotingPointForecaster using mean, weighted, and median aggregation strategies.