Faithful and Interpretable Explanations for Complex Ensemble Time Series Forecasts using Surrogate Models and Forecastability Analysis

2025-10-09

AI Summary

AutoML-based ensemble time-series forecasting models suffer from poor interpretability. Method: This paper proposes a dual-track interpretability framework integrating surrogate modeling and predictability analysis. First, a high-fidelity LightGBM surrogate model is trained to approximate the black-box AutoML model (e.g., AutoGluon), with local feature attributions derived via SHAP. Second, spectral predictability analysis quantifies intrinsic series predictability by comparing spectral density estimates against a noise baseline, serving as a confidence metric for explanations. Third, a per-item normalization strategy accommodates multi-scale, heterogeneous time series. Results: Experiments on the M5 dataset show that spectral predictability is significantly positively correlated with both forecasting accuracy and surrogate fidelity. The framework effectively identifies low-confidence predictions, enhances explanation stability, and improves user trust calibration.

Technology Category

• Machine Learning: Calibration & Uncertainty Quantification
• Natural Language Processing: Interpretability, Analysis, and Evaluation of NLP Models
• Planning, Routing, and Scheduling: Model-Based Reasoning

Application Category

• User Modeling, Personalization and Recommendation: Explainable and interpretable methods for personalization
• Web Mining and Content Analysis: Large pretrained models with web data
• Search and Retrieval-Augmented AI: Web learning to rank, online learning, and counterfactual learning for ranking

Abstract

Modern time series forecasting increasingly relies on complex ensemble models generated by AutoML systems like AutoGluon, delivering superior accuracy but with significant costs to transparency and interpretability. This paper introduces a comprehensive, dual-approach framework that addresses both the explainability and forecastability challenges in complex time series ensembles. First, we develop a surrogate-based explanation methodology that bridges the accuracy-interpretability gap by training a LightGBM model to faithfully mimic AutoGluon's time series forecasts, enabling stable SHAP-based feature attributions. We rigorously validated this approach through feature injection experiments, demonstrating remarkably high faithfulness between extracted SHAP values and known ground truth effects. Second, we integrated spectral predictability analysis to quantify each series' inherent forecastability. By comparing each time series' spectral predictability to its pure noise benchmarks, we established an objective mechanism to gauge confidence in forecasts and their explanations. Our empirical evaluation on the M5 dataset found that higher spectral predictability strongly correlates not only with improved forecast accuracy but also with higher fidelity between the surrogate and the original forecasting model. These forecastability metrics serve as effective filtering mechanisms and confidence scores, enabling users to calibrate their trust in both the forecasts and their explanations. We further demonstrated that per-item normalization is essential for generating meaningful SHAP explanations across heterogeneous time series with varying scales. The resulting framework delivers interpretable, instance-level explanations for state-of-the-art ensemble forecasts, while equipping users with forecastability metrics that serve as reliability indicators for both predictions and their explanations.

Problem

• Explaining complex ensemble time series forecasts for transparency and interpretability
• Quantifying time series forecastability to gauge prediction confidence
• Generating faithful explanations across heterogeneous time series with varying scales

Innovation

• Surrogate LightGBM model mimics AutoGluon forecasts
• Spectral predictability analysis quantifies series forecastability
• Per-item normalization enables meaningful SHAP explanations