20.500.12556/DiRROS-27644 Variational oblique predictive clustering trees Oblique predictive clustering trees (SPYCTs) are semi-supervised multi-target prediction models mainly used for structured output prediction (SOP) problems. They are computationally efficient and when combined in ensembles they achieve state-of-the-art results. However, one major issue is that it is challenging to interpret an ensemble of SPYCTs without the use of a model-agnostic method. We propose variational oblique predictive clustering trees, which address this challenge. The parameters of each split node are treated as random variables, described with a probability distribution, and they are learned through the Variational Bayes method. We evaluate the model on several benchmark datasets of different sizes. The experimental analyses show that a single variational oblique predictive clustering tree (VSPYCT) achieves competitive, and sometimes better predictive performance than the ensemble of standard SPYCTs. We also present a method for extracting feature importance scores from the model. Finally, we present a method to visually interpret the model’s decision making process through analysis of the relative feature importance in each split node. machine learning predictive clustering interpretable models structured output prediction uncertainty quantification napovedno gručenje razložljivi modeli napovedovanje strukturiranih izhodov kvantifikacija negotovosti true false true Elsevier Nizozemska Angleški jezik Slovenski jezik © 2026 The Author(s). Neznano 2026-02-17 11:39:54 2026-02-17 11:39:55 2026-02-18 03:58:03 0000-00-00 00:00:00 2026 0 0 Nasl. z nasl. zaslona; Soavtorji: Sašo Džeroski, Biljana Mileva Boshkoska, Pavle Boškoski; Opis vira z dne 28. 1. 2026; str. 1-31 Vol. [article no.] 131255 [in press] 2026 0000-00-00 Zaloznikova Objavljeno NiDoloceno 2025-04-29 2026-01-15 2026-01-22 004.8 1873-6793 10.1016/j.eswa.2026.131255 266512899 https://www.sciencedirect.com/science/article/pii/S0957417426001685 1 007cce4a-0bed-11f1-bcbc-001a4af901a5 https://dirros.openscience.si/Dokument.php?lang=slv&id=40549 RAZ_Andonovikj_Viktor_2026.pdf RAZ_Andonovikj_Viktor_2026.pdf 1 B7052A9221ABA62354203C8B8EAD7BE2 551974f2599bf6e4a085a5b43eeae0119d358ce15f84b92adeecbe2f4d43efbc 48036b6c-0bed-11f1-bcbc-001a4af901a5 https://dirros.openscience.si/Dokument.php?lang=slv&id=40550 Institut Jožef Stefan 0 0 0