Abstract

Abstract
Time series forecasting is pivotal across industries, as it fosters data-driven decision-making, increasing the chances of successful outcomes. Yet, certain instances that feature adverse characteristics, may lead models to manifest stress through decreases in performance (e.g., large errors). Hence, the ability to preemptively identify such cases, while establishing their root causes, would be advantageous to elevate the understanding of forecasting processes, informing users about the trustworthiness of predictions. Hence, we propose MASTFM, a method based on meta-learning that leverages statistical characteristics of input time series, and estimations of forecasting performance from model outputs, to build a metamodel that learns conditions for stress. Given that such occurrences are naturally rare, data augmentation is employed to ensure balance during training. Moreover, SHapley Additive exPlanations (SHAP) are used to explain how features impact forecasting behaviour.

Abstract
Time series forecasting is an important tool for planning and decision-making. Considering this, several forecasting algorithms can be used, with results depending on the characteristics of the time series. The recommendation of the most suitable algorithm is a frequent concern. Metalearning has been successfully used to recommend the best algorithm for a time series analysis task. Additionally, it has been shown that decomposition methods can lead to better results. Based on previously published studies, in the experiments carried out, time series components were used. This work proposes and empirically evaluates METAFORE, a new time series forecasting approach that uses seasonal trend decomposition with Loess and metalearning to recommend suitable algorithms for time series forecasting combinations. Experimental results show that METAFORE can obtain a better predictive performance than single models with statistical significance. In the experiments, METAFORE also outperformed models widely used in the state-of-the-art, such as the long short-term memory neural network architectures, in more than 70%\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$70\%$$\end{document} of the time series tested. Finally, the results show that the joint use of metalearning and time series decomposition provides a competitive approach to time series forecasting.

Abstract

Abstract

Abstract
Subgroup discovery aims to identify interpretable segments of a dataset where model behavior deviates from global trends. Traditionally, this involves uncovering patterns among data instances with respect to a target property, such as class labels or performance metrics. For example, classification accuracy can highlight subpopulations where models perform unusually well or poorly. While effective for model auditing and failure analysis, accuracy alone provides a limited view, as it does not reflect model confidence or sources of uncertainty. This work proposes a complementary approach: subgroup discovery using model uncertainty. Rather than identifying where the model fails, we focus on where it is systematically uncertain, even when predictions are correct. Such uncertainty may arise from intrinsic data ambiguity (aleatoric) or poor data representation in training (epistemic). It can highlight areas of the input space where the model's predictions are less robust or reliable. We evaluate the feasibility of this approach through controlled experiments on the classification of synthetic data and the Iris dataset. While our findings are exploratory and qualitative, they suggest that uncertainty-based subgroup discovery may uncover interpretable regions of interest, providing a promising direction for model auditing and analysis.