Model Evaluation and Visualization

Evaluating the performance of forecasting models and effectively visualizing their outputs are critical steps in the modeling lifecycle. fusionlab-learn offers tools to help you assess your models quantitatively and gain qualitative insights through plots.

This section covers:

• Evaluation Plotting: Demonstrations of how to use dedicated

  plotting functions to visualize forecast accuracy, uncertainty, and other performance aspects.

• Metrics: Detailed explanations of various metrics available for

  evaluating point forecasts, probabilistic (quantile or ensemble) forecasts, and other model characteristics like stability and calibration.

Dive into the following pages to understand how to rigorously evaluate and interpret your fusionlab-learn models.

Evaluation:

• Evaluating and Visualizing Forecasts
  • Prerequisites
  • Common Setup for Examples
  • 1. Visualizing Forecast Comparisons (plot_forecast_comparison)
  • 2. Visualizing Metrics Over the Forecast Horizon (plot_metric_over_horizon)
  • 3. Visualizing Metrics Across Segments with Radar Plots (plot_metric_radar)
  • Further Exploration
• Metrics for Forecasting Evaluation
  • Forecasting Metrics (fusionlab.metrics)