{"abstract":[{"lang":"eng","text":"While shallow decision trees may be interpretable, larger ensemble models like gradient-boosted trees, which often set the state of the art in machine learning problems involving tabular data, still remain black box models. As a remedy, the Shapley value (SV) is a well-known concept in explainable artificial intelligence (XAI) research for quantifying additive feature attributions of predictions. The model-specific TreeSHAP methodology solves the exponential complexity for retrieving exact SVs from tree-based models. Expanding beyond individual feature attribution, Shapley interactions reveal the impact of intricate feature interactions of any order. In this work, we present TreeSHAP-IQ, an efficient method to compute any-order additive Shapley interactions for predictions of tree-based models. TreeSHAP-IQ is supported by a mathematical framework that exploits polynomial arithmetic to compute the interaction scores in a single recursive traversal of the tree, akin to Linear TreeSHAP. We apply TreeSHAP-IQ on state-of-the-art tree ensembles and explore interactions on well-established benchmark datasets."}],"title":"Beyond TreeSHAP: Efficient Computation of Any-Order Shapley Interactions for Tree Ensembles","date_updated":"2024-03-27T15:06:39Z","issue":"13","intvolume":"        38","department":[{"_id":"660"}],"project":[{"_id":"126","name":"TRR 318 - C3: TRR 318 - Subproject C3"},{"_id":"109","grant_number":"438445824","name":"TRR 318: TRR 318 - Erklärbarkeit konstruieren"},{"_id":"117","name":"TRR 318 - C: TRR 318 - Project Area C"}],"volume":38,"publication_identifier":{"issn":["2374-3468","2159-5399"]},"author":[{"first_name":"Maximilian","last_name":"Muschalik","full_name":"Muschalik, Maximilian"},{"last_name":"Fumagalli","full_name":"Fumagalli, Fabian","id":"93420","first_name":"Fabian"},{"full_name":"Hammer, Barbara","last_name":"Hammer","first_name":"Barbara"},{"last_name":"Huellermeier","full_name":"Huellermeier, Eyke","first_name":"Eyke","id":"48129"}],"publication_status":"published","doi":"10.1609/aaai.v38i13.29352","date_created":"2024-03-27T14:50:04Z","publication":"Proceedings of the AAAI Conference on Artificial Intelligence","page":"14388-14396","language":[{"iso":"eng"}],"citation":{"bibtex":"@article{Muschalik_Fumagalli_Hammer_Huellermeier_2024, title={Beyond TreeSHAP: Efficient Computation of Any-Order Shapley Interactions for Tree Ensembles}, volume={38}, DOI={<a href=\"https://doi.org/10.1609/aaai.v38i13.29352\">10.1609/aaai.v38i13.29352</a>}, number={13}, journal={Proceedings of the AAAI Conference on Artificial Intelligence}, publisher={Association for the Advancement of Artificial Intelligence (AAAI)}, author={Muschalik, Maximilian and Fumagalli, Fabian and Hammer, Barbara and Huellermeier, Eyke}, year={2024}, pages={14388–14396} }","ieee":"M. Muschalik, F. Fumagalli, B. Hammer, and E. Huellermeier, “Beyond TreeSHAP: Efficient Computation of Any-Order Shapley Interactions for Tree Ensembles,” <i>Proceedings of the AAAI Conference on Artificial Intelligence</i>, vol. 38, no. 13, pp. 14388–14396, 2024, doi: <a href=\"https://doi.org/10.1609/aaai.v38i13.29352\">10.1609/aaai.v38i13.29352</a>.","apa":"Muschalik, M., Fumagalli, F., Hammer, B., &#38; Huellermeier, E. (2024). Beyond TreeSHAP: Efficient Computation of Any-Order Shapley Interactions for Tree Ensembles. <i>Proceedings of the AAAI Conference on Artificial Intelligence</i>, <i>38</i>(13), 14388–14396. <a href=\"https://doi.org/10.1609/aaai.v38i13.29352\">https://doi.org/10.1609/aaai.v38i13.29352</a>","ama":"Muschalik M, Fumagalli F, Hammer B, Huellermeier E. Beyond TreeSHAP: Efficient Computation of Any-Order Shapley Interactions for Tree Ensembles. <i>Proceedings of the AAAI Conference on Artificial Intelligence</i>. 2024;38(13):14388-14396. doi:<a href=\"https://doi.org/10.1609/aaai.v38i13.29352\">10.1609/aaai.v38i13.29352</a>","short":"M. Muschalik, F. Fumagalli, B. Hammer, E. Huellermeier, Proceedings of the AAAI Conference on Artificial Intelligence 38 (2024) 14388–14396.","mla":"Muschalik, Maximilian, et al. “Beyond TreeSHAP: Efficient Computation of Any-Order Shapley Interactions for Tree Ensembles.” <i>Proceedings of the AAAI Conference on Artificial Intelligence</i>, vol. 38, no. 13, Association for the Advancement of Artificial Intelligence (AAAI), 2024, pp. 14388–96, doi:<a href=\"https://doi.org/10.1609/aaai.v38i13.29352\">10.1609/aaai.v38i13.29352</a>.","chicago":"Muschalik, Maximilian, Fabian Fumagalli, Barbara Hammer, and Eyke Huellermeier. “Beyond TreeSHAP: Efficient Computation of Any-Order Shapley Interactions for Tree Ensembles.” <i>Proceedings of the AAAI Conference on Artificial Intelligence</i> 38, no. 13 (2024): 14388–96. <a href=\"https://doi.org/10.1609/aaai.v38i13.29352\">https://doi.org/10.1609/aaai.v38i13.29352</a>."},"keyword":["Explainable Artificial Intelligence"],"type":"journal_article","publisher":"Association for the Advancement of Artificial Intelligence (AAAI)","user_id":"93420","year":"2024","status":"public","_id":"53073","article_type":"original"}