Shapley Values for Feature Selection: The Good, the Bad, and the Axioms
Open accessThe DOI matches its Crossref record; 63 of the 64 references it cites were found in a registry or an index.
Abstract
The Shapley value has become popular in the Explainable AI (XAI) literature, thanks, to a large extent, to a solid theoretical foundation, including four “favourable and fair” axioms for attribution in transferable utility games. The Shapley value is provably the only solution concept satisfying these axioms. In this paper, we introduce the Shapley value and draw attention to its recent uses as a feature selection tool. We call into question this use of the Shapley value, using simple, abstract “toy” counterexamples to illustrate that the axioms may work against the goals of feature selection. From this, we develop a number of insights that are then investigated in concrete simulation settings, with a variety of Shapley value formulations, including SHapley Additive exPlanations (SHAP) and Shapley Additive Global importancE (SAGE). The aim is not to encourage any use of the Shapley value for feature selection, but we aim to clarify various limitations around their current use in the literature. In so doing, we hope to help demystify certain aspects of the Shapley value axioms that are viewed as “favourable”. In particular, we wish to highlight that the favourability of the axioms depends non-trivially on the way in which the Shapley value is appropriated in the XAI application.
Overview
The claims and statistics of this paper haven’t been checked yet.
Free: it reads the open-access text and checks the first 5 claims that cite a source we can check. Create a free account to check every claim.
Integrity signals
Registry
Cites
Statistics
Verification
Claims the paper makes
The claims and statistics of this paper haven’t been checked yet.
Free: it reads the open-access text and checks the first 5 claims that cite a source we can check. Create a free account to check every claim.
Claims other papers make about this one
Each is checked against this paper’s own text: does it say what the citing paper says it does?
The cited source says what the claim attributes to it.
In the citing paper
“we incorporated the use of interpretability tools (eg, SHAP 28 values, feature importance) to ensure transparency, a priority emphasized in recent ML reporting guidelines.”
From Linguistic markers for identifying post-traumatic stress disorder and associated symptoms: a systematic literature review (2025) · cites this paper as [28]
The check wasn’t confident either way, so it gives no verdict — never disputed.
In the citing paper
“To align with best practices 27 for systematic reviews of ML studies, we augmented PROBAST with criteria tailored to these challenges”
From Linguistic markers for identifying post-traumatic stress disorder and associated symptoms: a systematic literature review (2025) · cites this paper as [28]
Experimental — accuracy not yet measured
Self-check
Percentages, sums and reported differences re-computed from the paper’s own numbers.
The self-check needs the paper’s full text.
Citations
Cites
A Unified Approach to Interpreting Model Predictions
Lundberg, S., & Lee, S.-I. (2017) · arXiv · doi:10.48550/arxiv.1705.07874 · Cited by 23,707
Gauriat, C.-M., Pencolé, Y., Ribot, P., & Brouillet, G. (2024) · HAL (Le Centre pour la Communication Scientifique Directe) · doi:10.4230/oasics.dx.2024.27 · Cited by 12,586
From local explanations to global understanding with explainable AI for trees
Lundberg, S. M., Erion, G., Chen, H., DeGrave, A., Prutkin, J. M., Nair, B., Katz, R., Himmelfarb, J., Bansal, N., & Lee, S.-I. (2020) · Nature Machine Intelligence · doi:10.1038/s42256-019-0138-9 · Cited by 10,551
17. A Value for n-Person Games
Shapley, L. S. (1953) · Contributions to the Theory of Games (AM-28), Volume II · doi:10.1515/9781400881970-018 · Cited by 6,950
A survey on feature selection methods
Chandrashekar, G., & Sahin, F. (2014) · Computers & Electrical Engineering · doi:10.1016/j.compeleceng.2013.11.024 · Cited by 5,189
Shapley, L. S. (1988) · The Shapley Value · doi:10.1017/cbo9780511528446.003 · Cited by 4,954
Explaining prediction models and individual predictions with feature contributions
Štrumbelj, E., & Kononenko, I. (2014) · Knowledge and Information Systems · doi:10.1007/s10115-013-0679-x · Cited by 2,286
Tukey, J. W. (1962) · The Annals of Mathematical Statistics · doi:10.1214/aoms/1177704711 · Cited by 1,860
Consistent Individualized Feature Attribution for Tree Ensembles
Lundberg, S. M., Erion, G. G., & Lee, S.-I. (2018) · arXiv · doi:10.48550/arxiv.1802.03888 · Cited by 1,430
Budescu, D. V. (1993) · Psychological Bulletin · doi:10.1037/0033-2909.114.3.542 · Cited by 1,193
Research works printed but not linked: 1
- 18. Guyon, U. V. Luxburg, S. Bengio, H. Wallach, R. Fergus, S. Vishwanathan, and R. Garnett, Eds., vol. 30. Curran Associates, Inc., 2017.
Cited by
Items 101–200 of 347
Manandhar, P., Rafiq, H., Rodríguez-Ubiñas, E., Barbosa, J. D., Qureshi, O. A., Tarek, M., & Sgouridis, S. P. (2022) · Energies · doi:10.3390/en16010285 · Cited by 11
Xu, K., Long, D., Zhang, M., & Wang, Y. (2024) · Scientific Reports · doi:10.1038/s41598-024-80294-5 · Cited by 13
“…rules, providing a globally interpretable method based on optimal Shapley values to measure the contribution of each feature in the model 38 , and it demonstrates interpretable predictions for machine learning classifiers, addressing deficiencies caused by lack of feature directionality 39 .”
Future-Proofing EU-27 Energy Policies with AI: Analyzing and Forecasting Fossil Fuel Trends
Tudor, C. D., Şova, R., Stamatiou, P., Vlachos, V., & Polychronidou, P. (2025) · Electronics · doi:10.3390/electronics14030631 · Cited by 16
Cho, S.-H., & Kim, Y. (2025) · Journal of Clinical Medicine · doi:10.3390/jcm14061843 · Cited by 12
“To assess variable importance in the XGBoost model, SHAP values were calculated, as outlined in previous research [15,29].”
Nancy Ambritta, P., Mahalle, P. N., Bhapkar, H. R., Shinde, G. R., & Sable, N. P. (2025) · International Journal of Information Technology · doi:10.1007/s41870-025-02444-w · Cited by 11
Liu, L., Liang, W., Zhao, G., & Wu, P. (2025) · Tunnelling and Underground Space Technology · doi:10.1016/j.tust.2025.106678 · Cited by 10
Yin, A., Chen, X., He, H., Morris, A. P., Yuan, Q., Ma, X., & Yang, Z. (2025) · Travel Behaviour and Society · doi:10.1016/j.tbs.2025.101075 · Cited by 16
Assessing handwriting task difficulty levels through kinematic features: a deep-learning approach
Babushkin, V., Alsuradi, H., Jamil, M. H., Al-Khalil, M. O., & Eid, M. A. (2023) · Frontiers in Robotics and AI · doi:10.3389/frobt.2023.1193388 · Cited by 10
“and the weights are ω(S) = |S|!(|F|−|S|−1)! |F| ! (Fryer et al., 2021).”
BackgroundInterpretable AI/ML for High-stakes Tasks with Human-in-the-loop: Critical Review and Future Trends
Kovalerchuk, B. (2024) · doi:10.21203/rs.3.rs-3989807/v1 · Cited by 9
Driving the environmental extra mile – Car sharing and voluntary carbon dioxide offsetting
Haase, E. (2022) · Transportation Research Part D Transport and Environment · doi:10.1016/j.trd.2022.103361 · Cited by 9
Mvita, M. J., Zulu, N. G., & Thethwayo, B. M. (2025) · Engineering Applications of Artificial Intelligence · doi:10.1016/j.engappai.2025.111457 · Cited by 9
Huang, Y., Ding, L., Liu, W., Niu, H., Yang, M., Lyu, G., Lin, S., & Hu, Q. (2023) · Water · doi:10.3390/w15071326 · Cited by 10
Liver Disease Classification - An XAI Approach to Biomedical AI
Agbozo, E., & Balungu, D. M. (2024) · Informatica · doi:10.31449/inf.v48i1.4611 · Cited by 13
Ta, S., Alizadeh, S., Lakshminarayanan, S., & Ray, A. K. (2023) · The Canadian Journal of Chemical Engineering · doi:10.1002/cjce.25076 · Cited by 10
Huber, F., & Steinhage, V. (2024) · Geomatics · doi:10.3390/geomatics4030016 · Cited by 10
“While the four mentioned works [45–48] regarding this topic focus on defining a value function for Shapley values that will cause negative examples when applied as a tool for feature selection, we take learnings from their counterexamples and extract necessary conditions for defining the Shapley…”
BackgroundTrusting AI made decisions in healthcare by making them explainable
Žlahtič, B., Završnik, J., Kokol, P., Vošner, H. B., Sobotkiewicz, N., Schaubach, B. A., & Kirbiš, S. (2024) · Science in progress · doi:10.1177/00368504241266573 · Cited by 7
“The decision tree visualizations, a key aspect of our methodology, provide an intuitive and holistic understanding of decision pathways, which is often more aligned with the practical needs and interpretative habits of medical practitioners compared to the more fragmented insights offered by SHAP and LIME.”
BackgroundInfluentialRani, P., Pancholi, S., Shaw, V., Pandey, S., Atzori, M., & Kumar, S. (2024) · IEEE Transactions on Medical Robotics and Bionics · doi:10.1109/tmrb.2024.3504007 · Cited by 9
“Fryer et al. [16] using Shapley value from SHAP as a feature selection tool for the game-theoretic formation.”
Habbab, F. Z., Kampouridis, M., & Papastylianou, T. (2025) · Artificial Intelligence Review · doi:10.1007/s10462-024-11037-1 · Cited by 10
“Both SHAP and SAGE build on the concept of Shapley values [73]; in traditional co-operative game theory, Shapley values reflect a partitioning of the overall output of a group (or ‘grand-coalition’), which expresses this output as the sum of the individual contributions of its members, obtained by…”
Mata, L. A., Sousa, M., Vieira, P. A., Queluz, M. P., & Rodrigues, A. (2025) · IEEE Access · doi:10.1109/access.2025.3536281 · Cited by 11
“This library implements the calculation of the Shapley values [45] representing the significance of each model feature, at each observation.”
Farzanegan, B., & Jagannathan, S. (2025) · IEEE Transactions on Automation Science and Engineering · doi:10.1109/tase.2025.3554431 · Cited by 8
“The Shapley value-based approach has emerged as a widely used method for attribution of interpretable characteristics in machine learning models, promising for the advancement of explainable AI (XAI) [32].”
Bora, B. J., & Sharma, P. (2026) · Energy Conversion and Management · doi:10.1016/j.enconman.2025.121014 · Cited by 8
The disagreement dilemma in explainable AI: can bias reduction bridge the gap
Bhardwaj, N., & Parashar, G. (2025) · International Journal of Systems Assurance Engineering and Management · doi:10.1007/s13198-025-02712-9 · Cited by 8
“SHAP (Fryer et al. 2021) offers a comprehensive perspective on feature importance.”
Toward interpretable expected goals modeling using Bayesian mixed models
Iapteff, L., Coz, S. L., Rioland, M., Houde, T., Carling, C., & Imbach, F. (2025) · Frontiers in Sports and Active Living · doi:10.3389/fspor.2025.1504362 · Cited by 9
AI-Driven Optimization of Multilayer Thin-Film Structures for Advanced Optical Applications
Alkhazraji, E., Mukhtar, S., Khaled, H., & Viegas, J. (2025) · IEEE Access · doi:10.1109/access.2025.3591001 · Cited by 9
“SHAP values satisfy important axioms such as local accuracy, consistency, and missingness, making them ideal for attributing prediction influence to individual design variables [27].”
Kost, L., Lier, S. K., & Breitner, M. H. (2025) · Energy and AI · doi:10.1016/j.egyai.2025.100648 · Cited by 13
Ding, H., Zhao, J., Wang, W., Ding, R., Yun, Y., Liu, Y., Tu, K., Lan, W., & Pan, L. (2026) · Food Control · doi:10.1016/j.foodcont.2026.112026 · Cited by 7
Explainability of machine learning models under missing data
Vo, T. L., Nguyen, T. V., López-Ramos, L. M., Hammer, H. L., Riegler, M. A., & Halvorsen, P. (2026) · Applied Soft Computing · doi:10.1016/j.asoc.2026.115105 · Cited by 20
Singh, S. K., & Chaturvedi, A. (2024) · Engineering Applications of Artificial Intelligence · doi:10.1016/j.engappai.2024.109225 · Cited by 11
Analyzing Arabic Handwriting Style through Hand Kinematics
Babushkin, V., Alsuradi, H., Al-Khalil, M. O., & Eid, M. A. (2024) · Sensors · doi:10.3390/s24196357 · Cited by 7
“| F | ! are the weights [42].”
BackgroundMethodInferring feature importance with uncertainties with application to large genotype data
Johnsen, P. V., Strümke, I., Langaas, M., DeWan, A. T., & Riemer–Sørensen, S. (2023) · PLoS Computational Biology · doi:10.1371/journal.pcbi.1010963 · Cited by 14
“The SAGE framework thus provides an explanation of the influence of the features taking into account not only the model, but also implicitly the data via the loss function, thus encapsulating that the model might not be—and most likely is not—a perfect description of the data [18].”
On the Connection between Game-Theoretic Feature Attributions and Counterfactual Explanations
Albini, E., Sharma, S., Mishra, S., Dervovic, D., & Magazzeni, D. (2023) · Proceedings of the AAAI/ACM Conference on AI Ethics and Society · doi:10.1145/3600211.3604676 · Cited by 11
Measuring the Driving Forces of Predictive Performance: Application to Credit Scoring
Hué, S., Hurlin, C., Pérignon, C., & Saurin, S. (2026) · Management Science · doi:10.1287/mnsc.2023.02025 · Cited by 18
A comparison of local explanation methods for high-dimensional industrial data: A simulation study
Fries, N., & Rydén, P. (2022) · Expert Systems with Applications · doi:10.1016/j.eswa.2022.117918 · Cited by 9
Ho, T. K., Luo, Y.-F., & Guido, R. C. (2022) · IEEE Signal Processing Magazine · doi:10.1109/msp.2022.3155906 · Cited by 9
Deep Fusion Intelligence: Enhancing 5G Security Against Over-the-Air Attacks
Amini, M., Asemian, G., Kantarcı, B., Ellement, C., & Erol‐Kantarci, M. (2025) · IEEE Transactions on Machine Learning in Communications and Networking · doi:10.1109/tmlcn.2025.3533427 · Cited by 6
“Finally, to illustrate how predictors in the ensemble structure contribute to the classification output, Shapley values [46] are plotted in Fig.”
Ye, Z., Mahdad, A. T., Wang, Y., Shi, C., Chen, Y. J., & Saxena, N. (2025) · IEEE Symposium on Security and Privacy (S&P) · doi:10.1109/sp61157.2025.00049 · Cited by 10
“We then utilize Shapley value as a feature selection tool [51].”
Han, F., Lu, M., Qin, D., Zheng, G., Zeng, G., Yan, T., Wu, Z., Lu, H., Wang, J., Deng, Y., et al. (2025) · City and Environment Interactions · doi:10.1016/j.cacint.2025.100223 · Cited by 7
Paramasivam, P., Kanti, P. K., Dubey, A., Balasubramanian, D., Kale, U., & Kilikevičius, A. (2025) · International Journal of Hydrogen Energy · doi:10.1016/j.ijhydene.2025.150585 · Cited by 6
Explainable Anomaly Detection in Network Traffic Using Normalizing Flows
Shafir, L., Giryes, R., & Wool, A. (2025) · IEEE Transactions on Networking · doi:10.1109/ton.2025.3617580 · Cited by 7
“Furthermore, as discussed in some studies [36], the use of Shapley-based attribution methods for feature selection has been found to be unsuitable and, in certain cases, even counterproductive.”
An XGBoost-SHAP framework for interpretable and probabilistic flood susceptibility mapping
Rezvani, H., arfa, A., Shafizadeh‐Moghadam, H., & Minaei, M. (2026) · Natural Hazards · doi:10.1007/s11069-025-07908-7 · Cited by 8
“…a model’s prediction into the additive contribution of each feature, quantifying its marginal impact by calculating its average contribution across all possible feature combinations, thereby ensuring a fair allocation of the predictive "payout" among the input "players" (Fryer et al. 2021).”
Towards friendly AI: a comprehensive review and new perspectives on human-AI alignment
Sun, Q., Li, Y., Alturki, E., Murthy, S. M. K., & Schuller, B. W. (2026) · AI and Ethics · doi:10.1007/s43681-026-01070-x · Cited by 25
Nguyen, D., Chen, W., Guerrero‐Pérez, M. O., Rodríguez-Castellón, E., Nguyen, V. Q., Islam, A., Pham, P. P., & Hoang, A. T. (2026) · Applied Thermal Engineering · doi:10.1016/j.applthermaleng.2026.131703 · Cited by 8
Penny‐Dimri, J. C., Bergmeir, C., Reid, C. M., Williams‐Spence, J., Cochrane, A. D., & Smith, J. A. (2023) · PLoS ONE · doi:10.1371/journal.pone.0289930 · Cited by 6
Improving severity classification of Hebrew PET-CT pathology reports using test-time augmentation
Cohen, S., Lior, E., Bocher, M., & Rokach, L. (2023) · Journal of Biomedical Informatics · doi:10.1016/j.jbi.2023.104577 · Cited by 6
Textile Surface Defects Analysis with Explainable AI
Soon, R. J., & Chui, C. (2024) · IEEE Conference on Artificial Intelligence (CAI) · doi:10.1109/cai59869.2024.00248 · Cited by 6
Explainable Edge Computing in a Distributed AI - Powered Autonomous Vehicular Networks
Mahajan, P., Aujla, G. S., & Krishna, C. R. (2024) · IEEE International Conference on Communications workshops · doi:10.1109/iccworkshops59551.2024.10615824 · Cited by 9
“SHAP values can explain individual explanations and be applied to any AI model [20].”
Scherer, F. H., Dresch, F. N., Quincozes, S. E., Kreutz, D., & Quincozes, V. E. (2024) · Simpósio Brasileiro de Cibersegurança (SBSeg) · doi:10.5753/sbseg.2024.241780 · Cited by 6
Sun, C., & Liu, Z. (2024) · Computer Methods and Programs in Biomedicine · doi:10.1016/j.cmpb.2024.108481 · Cited by 10
Toumi, S., Aljadani, A., Toumi, H., Ammouri, B., & Dhiabi, M. (2025) · Eurasian economic review : · doi:10.1007/s40822-024-00295-7 · Cited by 10
“Feature selection is driven by various motivations, including the desire for computational efficiency in training and prediction, improved predictive power, and heightened interpretability (Guyon & Elisseeff, 2003; Fryer et al., 2021; Bhandari et al., 2020).”
Liu, X., & Aldrich, C. (2022) · Metals · doi:10.3390/met12111777 · Cited by 7
Temporal Analysis of Distribution Shifts in Malware Classification for Digital Forensics
Zola, F., Bruse, J. L., & Galar, M. (2023) · IEEE European Symposium on Security and Privacy Workshops (EuroS&PW) · doi:10.1109/eurospw59978.2023.00054 · Cited by 7
“However, in cases where ML decisions are complex and difficult to interpret, methods such as SHapley Additive exPlanation (SHAP) [16] or Local Interpretable Model-Agnostic Explanations (LIME) [47] can be employed to study the predictions of individual instances or cases (local interpretability) and…”
MethodWu, H., Badihi, H., Xue, Y., & Vilkko, M. K. (2025) · Sustainable Energy Grids and Networks · doi:10.1016/j.segan.2025.101775 · Cited by 10
Feature Selection for Battery Lifetime Prediction Using Explainable Machine Learning
Wang, Y., Wang, J., Chen, X., Wang, Y., & Song, J. (2025) · IEEE Transactions on Industry Applications · doi:10.1109/tia.2025.3618235 · Cited by 5
“Furthermore, our two-stage approach—combining a wrapper stage and an interpretability stage—aims to mitigate potential selection biases theoretically [16].”
Alyahya, A. N., Khan, S., Dilshad, N., & Ahmad, N. (2026) · Scientific Reports · doi:10.1038/s41598-026-41489-0 · Cited by 7
A Novel Random Forest Variant Based on Intervention Correlation Ratio
Zhang, T., Li, T., Xue, Z., Lu, X., & Gao, L. (2024) · IEEE Transactions on Emerging Topics in Computational Intelligence · doi:10.1109/tetci.2024.3369320 · Cited by 7
“This is consistent with the research results of [36] on Shap-ley characteristics concluding that the performance of Shapley value in feature selection is related to the application scenario.”
ResultExplainable paper classification system using topic modeling and SHAP
Shin, N., Lee, Y., Moon, H.-S., Kim, J., & Jung, H. (2024) · Intelligent Data Analysis · doi:10.3233/ida-240075 · Cited by 9
“Values close to zero imply minimal or negligible impact on the predictions [36,39].”
VMD-WSLSTM Load Prediction Model Based on Shapley Values
Shao, B., Yan, Y., & Zeng, H. (2022) · Energies · doi:10.3390/en15020487 · Cited by 5
“features and load compared to the traditional linear analysis method [26].”
MethodUkil, A., Marín, L., & Jara, A. J. (2022) · PLoS ONE · doi:10.1371/journal.pone.0277975 · Cited by 7
Ukil, A., Marín, L., & Jara, A. J. (2023) · IEEE International Conference on Acoustics Speech and Signal Processing · doi:10.1109/icassp49357.2023.10096437 · Cited by 5
“The classical Shapley value-based feature subset selection chooses k highest ranking ones [17], which demands some knowledge of k and the chance of sub-optimal selection is highly likely.”
BackgroundMethodExplainable Machine Learning for Categorical and Mixed Data with Lossless Visualization
Kovalerchuk, B., & McCoy, E. (2024) · Studies in computational intelligence · doi:10.1007/978-3-031-46549-9_3 · Cited by 8
An Explainable AI approach towards Epileptic Seizure Detection
Chapatwala, N., Paunwala, C. N., & Dalal, P. (2022) · IEEE India Conference · doi:10.1109/indicon56171.2022.10039982 · Cited by 5
“Shapley score does not prove directly for feature selection [19].”
BackgroundExplanation with the Winter value: Efficient computation for hierarchical Choquet integrals
Labreuche, C. (2022) · International Journal of Approximate Reasoning · doi:10.1016/j.ijar.2022.09.008 · Cited by 7
Khene, Z., Kutikov, A., Campi, R., & the EAU‐YAU Renal Cancer Working Group (2023) · British Journal of Urology · doi:10.1111/bju.16016 · Cited by 5
Explainable AI (XAI) for Agriculture
Linheiro, E. S. M., Shinde, G. R., Mahalle, P. N., & Mirajkar, R. (2023) · IoT and Big Data Analytics · doi:10.2174/9789815179187123040014 · Cited by 4
SVAP: Shapley Value Guided Attribution Prior for Neural Network-Based Autonomous Driving
Li, M., Sun, H., Huang, Y., & Chen, H. (2025) · IEEE Transactions on Vehicular Technology · doi:10.1109/tvt.2025.3575760 · Cited by 5
“For models trained on tabular data, feature attribution measures the individual contribution of each feature to the model’s output [24], [25], enabling us to evaluate the impact of each feature on the model’s prediction.”
Nopour, R. (2025) · Heliyon · doi:10.1016/j.heliyon.2025.e43510 · Cited by 10
Li, Y., Lu, M., Campos, L. C., Cheng, S., Li, Z., & Hu, Y. (2025) · Energy Nexus · doi:10.1016/j.nexus.2025.100583 · Cited by 4
Pham, D. T., Vujanović, M., Luu, V. C., Paramasivam, P., Nguyen, V. N., Efremov, C., Sănduleac, M., Nguyen, X. P., Trần, V. D., & Hoang, A. T. (2025) · Case Studies in Thermal Engineering · doi:10.1016/j.csite.2025.107337 · Cited by 4
Optimizing Intrusion Detection with Hybrid Deep Learning Models and Data Balancing Techniques
Sarwar, M. N., Arman, M. S., Bhuiyan, T., & Rafiq, F. B. (2025) · IEEE International Conference on AI in Cybersecurity (ICAIC) · doi:10.1109/icaic63015.2025.10849340 · Cited by 7
“SHAP values offer a game-theoretic approach to feature importance, aiding in feature selection based on their impact on predictions [25].”
Analyzing handwriting legibility through hand kinematics
Babushkin, V., Alsuradi, H., Al-Khalil, M. O., & Eid, M. A. (2025) · Frontiers in Artificial Intelligence · doi:10.3389/frai.2025.1426455 · Cited by 5
Understanding interpretable patterns of Shapley behaviours in materials data
Liu, T., & Barnard, A. S. (2025) · Machine Learning Engineering · doi:10.1088/3049-4761/adaaf6 · Cited by 6
Jiang, Y., Peng, R., Liu, X., Xu, M., Shen, H., Yu, Z., & Jiang, Z. (2025) · Academic Radiology · doi:10.1016/j.acra.2025.07.044 · Cited by 6
Detection of DDoS Attacks Using SHAP-Based Feature Reduction
Cynthia, C., Ghosh, D., & Kamath, G. K. (2023) · International Journal of Machine Learning · doi:10.18178/ijml.2023.13.4.1147 · Cited by 6
“SHapley Additive exPlanations (SHAP) [17] is one of the feature selection methods which provides feature importance.”
BackgroundSaliba, A., Tout, K. R., Zaki, C., & Claramunt, C. (2024) · Journal of Location Based Services · doi:10.1080/17489725.2023.2298803 · Cited by 4
“XGBoost uses regularisation techniques to reduce overfitting and provides feature importance metrics (Fryer, Strumke, and Nguyen 2021).”
MethodMitigating Extreme Cold Start in Graph-based RecSys through Re-ranking
Sbandi, A., Siciliano, F., & Silvestri, F. (2024) · ACM International Conference on Information and Knowledge Management (CIKM) · doi:10.1145/3627673.3680069 · Cited by 4
“In this situation, the features themselves become the players in a cooperative selection game, and model performance acts as the payoff [11, 15].”
Liu, L., Hong, Z.-X., Zhao, G., & Liang, W. (2024) · Journal of Central South University · doi:10.1007/s11771-024-5641-4 · Cited by 4
“In this study, the Shapley additive explanations (SHAP) method was adopted to reveal the relative importance of all input indicators [64 − 66].”
COFA: counterfactual attention framework for trustworthy wafer map failure classification
Feng, K., Wang, J., Yin, C., & Li, A. (2025) · Applied Intelligence · doi:10.1007/s10489-025-06488-0 · Cited by 4
Pimentel, J., McEwan, A. A., & Yu, H. Q. (2022) · Euromicro Conference on Digital System Design (DSD) · doi:10.1109/dsd57027.2022.00098 · Cited by 4
“SHAP (SHapley Additive exPlanation), a solution based on cooperative game theory, stands out as one of the most prominent XAI techniques [31].”
MethodOlsen, L. H. B., Glad, I. K., Jullum, M., & Aas, K. (2021) · arXiv (Cornell University) · doi:10.48550/arxiv.2111.13507 · Cited by 28
“The latter two are both Gaussian and can be evaluated at the specific vector x ∗S using the R -package mvtnorm , see Genz and Bretz (2009).”
MethodCan ESG Shape Cost of Capital? A Bibliometric Review and Empirical Analysis Through MLPreprint
Bussmann, N., Tanda, A., & Yu, E. P. (2022) · SSRN Electronic Journal · doi:10.2139/ssrn.4173890 · Cited by 4
Machine Learning-Based Approach for the Gambling Problem Identification
Kozák, J. T., Probierz, B., Juszczuk, P., Dziczkowski, G., Jach, T., Stefański, P., Głowania, S., Hrabia, A., Wołek, G., Sznapka, W., et al. (2025) · Vietnam Journal of Computer Science · doi:10.1142/s2196888825500034 · Cited by 3
Pixel-Based Clustering for Local Interpretable Model-Agnostic Explanations
Qian, J., Wen, T., Ling, M., Du, X., & Ding, H. (2025) · Journal of Artificial Intelligence and Soft Computing Research · doi:10.2478/jaiscr-2025-0013 · Cited by 3
A Comprehensive Study of Shapley Value in Data Analytics
Lin, H., Wan, S., Xie, Z., Chen, K., Zhang, M., Shou, L., & Chen, G. (2025) · Proceedings of the VLDB Endowment · doi:10.14778/3746405.3746429 · Cited by 11
“Test Accuracy/Error Rate of the ML model[3, 7, 9, 11, 23, 25, 30, 33, 34, 38, 41, 44, 48–50, 58, 63, 65, 67–69, 81, 83, 87, 89, 101, 107, 108, 113–115, 117, 120, 122, 127, 128, 132, 140, 143, 144, 147, 150, 151, 155, 156, 163, 164, 164, 167, 169].”
Alaoui, E. A. A., Sahlaoui, H., Sallah, A., & Nayyar, A. (2026) · Information Sciences · doi:10.1016/j.ins.2026.123128 · Cited by 5
Leveraging machine learning for digital gait analysis in ataxia using sensor-free motion capture
Wegner, P., Grobe‐Einsler, M., Reimer, L. M., Kahl, F., Koyak, B., Elter, T., Lange, A., Kimmich, O., Soub, D., Hufschmidt, F., et al. (2026) · Communications Medicine · doi:10.1038/s43856-025-01258-y · Cited by 4
“This offered interpretability of model predictions, in line with current explainable ML practices 33 .”
Xu, X., Wu, H., Yu, W., Hu, L., Jiang, P., & Gai, K. (2024) · International World Wide Web Conference (WWW) · doi:10.1145/3589335.3648345 · Cited by 4
“We interpret feature importance of each numerical feature using Shapley Value, which is one of the most widely adopted measures of feature importance as it has a solid theoretical foundation [7, 18].”
On many-objective feature selection and the need for interpretability
Njoku, U. F., Abelló, A., Bilalli, B., & Bontempi, G. (2024) · Expert Systems with Applications · doi:10.1016/j.eswa.2024.126191 · Cited by 3
Wu, S.-C. J., Sharma, N., Bauch, A., Yang, H.-C., Hect, J. L., Thomas, C., Wagner, S., Förstner, B. R., Arnim, C. A. F. V., Kaufmann, T., et al. (2025) · JMIR Aging · doi:10.2196/67958 · Cited by 4
“First, interpreting and using SHAP values warrants caution [67], as is generally the case for methods using model explanations in medicine [68,69].”
Tasdemir, E., ÇETİNKAYA, M. Y., Uysal, F., & El-Zahab, S. (2025) · Buildings · doi:10.3390/buildings15132307 · Cited by 4
Improving REITs Time Series Prediction Using ML and Technical Analysis Indicators
Habbab, F. Z., Kampouridis, M., & Papastylianou, T. (2023) · International Joint Conference on Neural Networks (IJCNN) · doi:10.1109/ijcnn54540.2023.10191969 · Cited by 3
“importance of the various features by using Shapley values, which is a commonly used tool for model explainability [33, 34].”
MethodXAI-Guided Optimization of a Multilayer Network Regression Model
Duszyńska, K., Polski, P., Włosek, M., Knapińska, A., Lechowicz, P., & Walkowiak, K. M. (2024) · IFIP Networking Conference (IFIP Networking) · doi:10.23919/ifipnetworking62109.2024.10619785 · Cited by 5
“In particular, previous studies showed the benefits of SHAP values for feature selection [32], [33].”
Mereological approach to higher-order structure in complex systems: From macro to micro with Möbius
Jansma, A. (2025) · Physical Review Research · doi:10.1103/physrevresearch.7.023016 · Cited by 14
Advancing Fact Attribution for Query Answering: Aggregate Queries and Novel Algorithms
Abramovich, O., Deutch, D., Frost, N., Kara, A., & Olteanu, D. (2025) · Proceedings of the VLDB Endowment · doi:10.14778/3749646.3749670 · Cited by 4
Ganesh, K., Chatterjee, A., Riegler, M. A., & Halvorsen, P. (2023) · Research Square · doi:10.21203/rs.3.rs-3114142/v4 · Cited by 3
Simple and Effective Portfolio Construction with Crypto AssetsPreprint
Johansson, K., & Boyd, S. (2024) · SSRN Electronic Journal · doi:10.2139/ssrn.5043017 · Cited by 7
Quillivic, R., Auxéméry, Y., Gayraud, F., Dayan, J., & Mesmoudi, S. (2025) · Journal of the American Medical Informatics Association · doi:10.1093/jamia/ocaf075 · Cited by 3
REFRESH: Responsible and Efficient Feature Reselection guided by SHAP values
Sharma, S., Dutta, S., Albini, E., Lécué, F., Magazzeni, D., & Veloso, M. (2023) · Proceedings of the AAAI/ACM Conference on AI Ethics and Society · doi:10.1145/3600211.3604706 · Cited by 4
Rapid noninvasive lung cancer screening via discriminative wavenumbers in Raman spectroscopy
Hano, H., Suárez, B., Amigo, J. M., Lawrie, C. H., & Seifert, A. (2025) · Microchemical Journal · doi:10.1016/j.microc.2024.112496 · Cited by 2
Boutaghane, A., Ameur-Zaimeche, O., Heddam, S., Kechiched, R., Tahar-Belkacem, N., Ouladmansour, A., Al‐Mudhafar, W. J., & Wood, D. A. (2025) · Earth Science Informatics · doi:10.1007/s12145-024-01642-7 · Cited by 2
“These results provide enhancement of black-box algorithms transparency and interpretability by attempting to explain how the model arrived at a certain prediction (Fryer et al. 2021).”
Semantic Shapley-based counterfactual explanations for end-to-end autonomous driving
Sun, H., Li, M., Cui, Z., Huang, Y., & Chen, H. (2025) · Engineering Applications of Artificial Intelligence · doi:10.1016/j.engappai.2025.111638 · Cited by 4
Metadata
- Authors
- Fryer, D., Strumke, I., & Nguyen, H.
- Affiliations
- The University of Queensland (AU)
- La Trobe University (AU)
- Venue
- IEEE Access
- Published
- 2021
- Volume
- 9
- Pages
- 144352-144360
- ISSN
- 2169-3536 (ISSN-L)
- OpenAlex
- W3131037858
- Semantic Scholar
- 6573fa74af238d1e3538c026997e31b9f67f19f7
- Cited by
- 347
- References
- 64
- Citations by year
- 2026: 56 · 2025: 102 · 2024: 86 · 2023: 58 · 2022: 34 · 2021: 11
- FWCI
- 16.63
- Type
- journal article
- Language
- en
- Open access
- gold
- doi.org
- Licence
- Locations
- IEEE Access (publishedVersion, cc-by)
- arXiv (Cornell University) (submittedVersion)
- DOAJ (DOAJ: Directory of Open Access Journals) (submittedVersion, cc-by-sa)
- Figshare (submittedVersion, cc-by)
- arXiv (Cornell University)
- La Trobe University (cc-by)
- Topics
- Physical Sciences › Computer Science › Artificial Intelligence › Explainable Artificial Intelligence (XAI)
- Social Sciences › Economics, Econometrics and Finance › Economics and Econometrics › Game Theory and Voting Systems
- Physical Sciences › Computer Science › Artificial Intelligence › Privacy-Preserving Technologies in Data
- Keywords
- Shapley value, feature selection, explainable AI, axioms, SHAP, sage, counterexample, transferable utility games, attribution
About these checks
Registry, retraction, Cites and statistics results are computed, not judged, with one exception: a printed reference that no index lists is read once by a model before it counts as a citation error. Claims are extracted by a model and kept only when their quote appears verbatim in the text; each cited source is then checked, with a model’s help, for whether it says what the claim relies on. Those verdicts can be wrong — read the claim’s quote and the cited source before relying on one.
Statistics and arithmetic are recomputed, not judged.
Confidence bands are shown only for fields calibrated on our annotated test set. None is calibrated yet, so no band is shown.
What counts as a claim ›How we measure confidence ›How we check this ›
Are you an author, or do you see a problem with this page? Request a review
Data sources
Bibliographic data: Semantic Scholar, OpenAlex, Crossref.
Contains information from the Semantic Scholar Academic Graph Datasets, made available under the ODC Attribution License (ODC-BY 1.0).
Semantic Scholar’s data is described in: Kinney, R., et al. (2023). The Semantic Scholar Open Data Platform. arXiv:2301.10140.