Breast Cancer Classification using Machine Learning Techniques with Explainable AI

All the symptoms have been analyzed using several machine learning algorithms for diagnosing breast cancer. This paper utilizes the Breast Cancer Wisconsin (Diagnostic) data set to show how individuals at risk of having malignant tumors can be classified into benign tumors and malignant tumors. Preprocessing phase involved methods like Feature selection, data scaling and managing class imbalances. Out of seven machine learning algorithms, we compared Logistic Regression, Support Vector Machines (SVM), Gradient Boosting Classifier, Random Forest, K-Nearest Neighbors (KNN), XGBoost, and Decision Tree. Evaluation criterion included accuracy, precision, recall, F1 score, and confusion matrix. SVM appears to be the most applicable model in medical data classification with a great accuracy of 98.25% than other models shown in the results above. This research uses two XAI methods, namely SHAP and LIME, to explain feature significance and improve the interpretability of the SVM model. The discussion of the tumor-related features in this work provides focus and guidance for clinical applications to improve detection and prevention of the disease in its preliminary stages to improve health care.