Explainable Machine Learning Framework for Breast Cancer Survival Prediction Using the METABRIC Dataset

Upadhyay, Ragini ORCID: https://orcid.org/0009-0001-3207-4501 (2026) Explainable Machine Learning Framework for Breast Cancer Survival Prediction Using the METABRIC Dataset. In: 11th International Conference on Computer and Information Science and Technology (CIST 2026), 16-18 August 2026, London. ISSN 2369-3010 (In Press)

Other (First-authored research paper presented at CIST 2026 in London. The study developed an explainable machine-learning framework for predicting breast cancer survival using the METABRIC dataset, incorporating SHAP and LIME to interpret model predictions a) CIST26-136 (1).docx - Published Version Restricted to Repository staff only (Author is actively seeking publication). Available under License All Rights Reserved. Download (432kB)

Text 16511 Upadhyay, R. (2026) Explainable Machine Learning Framework for Breast Cancer.pdf - Accepted Version Restricted to Repository staff only (Author is actively seeking publication). Available under License All Rights Reserved. Download (682kB)

Abstract

Breast cancer remains one of the leading causes of cancer-related mortality among women worldwide and continues to pose significant challenges to healthcare systems, including the United Kingdom National Health Service (NHS). Early diagnosis and accurate survival prediction are essential for improving treatment planning, personalised interventions, and patient outcomes. Recent advancements in artificial intelligence (AI) and machine learning (ML) have enabled the development of predictive models capable of analysing large-scale clinical and genomic datasets for cancer prognosis. However, many high-performing AI systems operate as black-box models, limiting clinical trust and interpretability. This study proposes an explainable machine learning framework for breast cancer survival prediction using the Molecular Taxonomy of Breast Cancer International Consortium (METABRIC) dataset. The framework integrates ensemble learning models — including Random Forest and Extreme Gradient Boosting (XGBoost) — with SHapley Additive explanations (SHAP) to provide both global and patient-specific interpretability. The proposed methodology involves data preprocessing, feature engineering, survival risk classification, model training, comparative evaluation, and explainability analysis. The results demonstrate that ensemble learning models combined with SHAP-based interpretability provide high predictive performance while improving transparency and usability in healthcare environments. The study contributes to the growing field of explainable AI in healthcare by bridging the gap between predictive accuracy and clinical transparency.

Item Type:	Conference or Workshop Item (Paper)
Uncontrolled Keywords:	Breast Cancer; Explainable Artificial Intelligence; Machine Learning; XGBoost; Random Forest, SHAP; METABRIC Dataset; Survival Prediction; Clinical Decision Support
Subjects:	Q Science > QA Mathematics
Divisions:	Schools and Research Institutes > School of Business, Computing and Social Sciences
Depositing User:	Ragini Upadhyay
Date Deposited:	21 Oct 2026 12:09
Last Modified:	21 Oct 2026 12:15
URI:	https://eprints.glos.ac.uk/id/eprint/16511

University Staff: Request a correction | Repository Editors: Update this record