EUVIMEDEUVIMED entdecken
Crossref

Investigating the impact of balancing, filtering, and complexity on predictive multiplicity: A data-centric perspective

Mustafa Cavus, Przemysław Biecek

Information Fusion · 2025-11 · 123 · 103243

DOI: 10.1016/j.inffus.2025.103243 ↗

Kurz gefasst

Fachartikel in der Zeitschrift Information Fusion, 2025-11. Verfasst von Mustafa Cavus und 1 weiteren Autor:innen. Die Arbeit wurde bisher 9-mal zitiert (Crossref) und nennt 78 Literaturangaben. Gefördert von: Eskişehir Technical University, National Science Centre Poland.

AbstractBibliografische AngabenZitierenVerwandte Studien

Abstract (Originalsprache)

The Rashomon effect presents a significant challenge in model selection. It occurs when multiple models achieve similar performance on a dataset but produce different predictions, resulting in predictive multiplicity. This is especially problematic in high-stakes environments, where arbitrary model outcomes can have serious consequences. Traditional model selection methods prioritize accuracy and fail to address this issue. Factors such as class imbalance and irrelevant variables further complicate the situation, making it harder for models to provide trustworthy predictions. Data-centric AI approaches can mitigate these problems by prioritizing data optimization, particularly through preprocessing techniques. However, recent studies suggest preprocessing methods may inadvertently inflate predictive multiplicity. This paper investigates how data preprocessing techniques like balancing and filtering methods impact predictive multiplicity and model stability, considering the complexity of the data. We conduct the experiments on 21 real-world datasets, applying various balancing and filtering techniques, and assess the level of predictive multiplicity introduced by these methods by leveraging the Rashomon effect. Additionally, we examine how filtering techniques reduce redundancy and enhance model generalization. The findings provide insights into the relationship between balancing methods, data complexity, and predictive multiplicity, demonstrating how data-centric AI strategies can improve model performance.

Förderung

  • Eskişehir Technical University
  • National Science Centre Poland

Zitieren

APA

Cavus, M., & Biecek, P. (2025). Investigating the impact of balancing, filtering, and complexity on predictive multiplicity: A data-centric perspective. Information Fusion, 123, 103243. https://doi.org/10.1016/j.inffus.2025.103243

Vancouver

Cavus M, Biecek P. Investigating the impact of balancing, filtering, and complexity on predictive multiplicity: A data-centric perspective. Information Fusion. 2025;123:103243. doi:10.1016/j.inffus.2025.103243

BibTeX
@article{Cavus2025,
  author = {Cavus, Mustafa and Biecek, Przemysław},
  title = {Investigating the impact of balancing, filtering, and complexity on predictive multiplicity: A data-centric perspective},
  journal = {Information Fusion},
  year = {2025},
  volume = {123},
  pages = {103243},
  doi = {10.1016/j.inffus.2025.103243}
}
Kostenlos starten →