Comparative Analysis of Explainable AI Methods Across Machine Learning Classifiers for Breast Cancer Detection

Read full text Summarise with VIRAT
Authors: Rajdeep Mondal, Rohan Kumar Mondal, Ansh Jamwal
Cited by: 0
References: 12

Summary

No summary yet. VIRAT can read this paper and write a plain-language summary of its aim, methods, key findings and limitations.

Summarise with VIRAT Free to try with your Science Catalogs account.

Abstract

Abstract Breast cancer has been one of the leading causes of death among women in the world and the importance of early and precise detection of the disease has been of utmost importance in the world. Machine learning models have been proven to be highly predictive when applied in medical diagnosis; their inability to offer interpretability can however act as a barrier when it comes to their use in clinical practice. Explainable Artificial Intelligence (XAI) systems, including SHAP (Shapley Additive Explanations) and LIME (Local Interpretable Model-agnostic Explanations) have become one of the most effective to combat the issue. This work provides a comparative study of SHAP and LIME on two popular classifiers- Logistic Regression and random forest on the Wisconsin Breast Cancer data set. Along with measuring the performance of a model, the paper presents a quantitative measure of agreement to evaluate the reliability of SHAP and LIME in the presence of important features. The experimental findings demonstrate that the two approaches reach a great degree of consensus (approximately 77 percent), which would lead to consistent and high-quality explanations, and slight deviations can be observed in more complicated models. These results indicate the complementary relationship between SHAP and LIME and their possibility in improving medical AI system transparency.

Fields of Study & Subjects

Explainable Artificial Intelligence (XAI) Artificial Intelligence in Healthcare AI in cancer detection

Publication details

Journal / source
Research Square
Publication date
25 March 2026
Language
EN
Document type
Preprint