Comparing Explainability Techniques for Recommender Systems: A Theoretical Perspective on Shapley Values

Paper ID: EIJTEM_2025_12_4_7-13

Author's Name: Drashti Shrimal, Dr. Harshali Patil

Volume: 12

Issue: 4

Year: 2025

Page No: 7-13

Abstract:

As recommendation systems become increasingly central to digital platforms, the need to explain their decisions has moved from a nice-to-have to a necessity. Users often receive recommendations without understanding why, and this lack of transparency can erode trust and hinder adoption. This paper explores the potential of Shapley values—a concept from cooperative game theory— as a robust and intuitive approach to explain recommendations. Unlike many existing interpretability techniques that rely on heuristics or model-specific mechanisms, Shapley values offer a theoretically grounded method to fairly attribute a model’s output to its input features. We present a conceptual framework that illustrates how Shapley- based techniques, including KernelSHAP, TreeSHAP, DeepSHAP, and TimeSHAP, can be applied across different types of recommendation models. Through a comparative theoretical lens, we examine the strengths, assumptions, and limitations of these techniques, and we argue why Shapley values hold unique promise for building transparent, user-aligned, and accountable recommendation systems. This work aims to bridge the gap between explainability theory and its application in real- world recommender contexts.

Keywords: Shapley values, explainable recommendation systems, model interpretability, KernelSHAP, transparency in AI

References:

1. Nohara, Y., Matsumoto, K., Soejima, H., & Nakashima, N. (2022). Explanation of machine learning models using shapley additive explanation and application for real data in hospital. Computer Methods and Programs in Biomedicine, 214, 106584.
2. Shapley, L. S. (1953). A value for n-person games.
3. Chen, H., Covert, I. C., Lundberg, S. M., & Lee,
S. I. (2023). Algorithms to estimate Shapley value feature attributions. Nature Machine Intelligence, 5(6), 590-601.
4. Chowdhury, K. R., Sil, A., & Shukla, S. R. (2021). Explaining a black-box sentiment analysis model with local interpretable model diagnostics explanation (LIME). In Advances in Computing and Data Sciences: 5th International Conference, ICACDS 2021, Nashik, India, April 23–24, 2021, Revised Selected Papers, Part I 5 (pp. 90-101). Springer International Publishing.
5. Daoud, M. S., Shehab, M., Al-Mimi, H. M., Abualigah, L., Zitar, R. A., & Shambour, M. K.
Y. (2023). Gradient-based optimizer (GBO): a review, theory, variants, and applications. Archives of Computational Methods in Engineering, 30(4), 2431-2449.
6. Tutek, M., & Šnajder, J. (2022). Toward practical usage of the attention mechanism as a tool for interpretability. IEEE access, 10, 47011-47030.
7. Roshan, K., & Zafar, A. (2022, March). Using kernel shap xai method to optimize the network anomaly detection model. In 2022 9th International Conference on Computing for Sustainable Global Development (INDIACom) (pp. 74-80). IEEE.
8. Yang, J. (2021). Fast treeshap: Accelerating shap value computation for trees. arXiv preprint arXiv:2109.09847.
9. Fernando, Z. T., Singh, J., & Anand, A. (2019, July). A study on the Interpretability of Neural Retrieval Models using DeepSHAP. In Proceedings of the 42nd international ACM SIGIR conference on research and development in information retrieval (pp. 1005-1008).
10. Bento, J., Saleiro, P., Cruz, A. F., Figueiredo, M. A., & Bizarro, P. (2021, August). Timeshap: Explaining recurrent models through sequence perturbations. In Proceedings of the 27th ACM SIGKDD conference on knowledge discovery & data mining (pp. 2565-2573).
11. Kumar, I. E., Venkatasubramanian, S., Scheidegger, C., & Friedler, S. (2020, November). Problems with Shapley-value- based explanations as feature importance measures. In International conference on machine learning (pp. 5491-5500). PMLR.
12. Huang, X., & Marques-Silva, J. (2024). On the failings of Shapley values for explainability. International Journal of Approximate Reasoning, 171, 109112.
13. Ge, Y., Tan, J., Zhu, Y., Xia, Y., Luo, J., Liu, S.,
... & Zhang, Y. (2022, July). Explainable fairness in recommendation. In Proceedings of the 45th International ACM SIGIR Conference on Research and Development in Information Retrieval (pp. 681-691).
14. Abdollahi, B., & Nasraoui, O. (2018). Transparency in fair machine learning: the case of explainable recommender systems. Human and machine learning: Visible, explainable, trustworthy and transparent, 21-35.
15. Vultureanu-Albişi, A., & Bădică, C. (2021, August). Recommender systems: An explainable AI perspective. In 2021 International conference on innovations in intelligent systems and applications (INISTA) (pp. 1-6). IEEE.

View PDF