Performance Analysis of Machine Learning Classifiers for Metastatic Breast Cancer Diagnosis Within 90 Days

Metastatic breast cancer, commonly referred to as stage IV breast cancer or advanced breast cancer, is a severe form of the disease, occurs when breast cancer cells migrate to different parts of the body such as the lungs, bones, liver and brain. This study focuses on predicting the of metastatic breast cancer diagnosis within a 90-day window. The dataset was meticulously processed, addressing missing values and encoding categorical variables to ensure a more structured and interpretable dataset. Visualizations were utilized to explore insights and relationships within the data. The results of different machine learning algorithms and deep learning model were evaluated, including Naive Bayes, Decision Tree, Random Forest, Logistic Regression, Support Vector Machine (SVM), Gradient Boosting Machine (GBM), and TabNet. Our findings reveal that GBM and TabNet outperformed other algorithms with respect to performance, achieving accuracies of 81.68% and 81.64%, respectively. Furthermore, each algorithm's performance was evaluated using the AUC-ROC curve.