Computer Science Faculty Publications
Document Type
Conference Proceeding
Publication Date
2025
Abstract
Time series data is one of the most ubiquitous data modality existing in a diverge critical domain such as healthcare, seismology, manufacturing and energy. Recent years, there are increasing interest of the data mining community to develop time series deep learning model to pursue better performance. The models performance often evaluate by certain evaluation metrics such as RMSE, Accuracy, and F1-score. Yet time series data are often hard to interpret and is collected with unknown environment factor, sensor configuration, latent physic mechanisms, and non-stationary evolving behavior. As a result, a model that is better on standard metric-based evaluation may not always perform better in the real-world tasks. In this blue sky paper, we aim to explore the challenge existed in the metric-based evaluation framework for time series data mining and propose a potential blue-sky idea — developing a knowledge-discovery-based evaluation framework, which aims to effectively utilize domain-expertise knowledge to evaluate model. We demonstrate that an evidence-seeking explanation can potentially has stronger persuasive power than metric-based evaluation and obtain better generalization ability for time series data mining tasks.
Recommended Citation
Zhang, Li. “Evaluating Time Series Models with Knowledge Discovery.” In Proceedings of the 2025 SIAM International Conference on Data Mining (SDM). Proceedings. Society for Industrial and Applied Mathematics, 2025. https://doi.org/10.1137/1.9781611978520.36.
Publication Title
Proceedings of the 2025 SIAM International Conference on Data Mining (SDM)
DOI
10.1137/1.9781611978520.36
Comments
Copyright © 2025 by SIAM Unauthorized reproduction of this article is prohibited