INTERNATIONAL JOURNAL FOR INNOVATIVE RESEARCH IN MULTIDISCIPLINARY FIELD

Abstract:  Accurate and early recognition of breast cancer significantly recovers survival rates and treatment outcomes. This research paper presents a comprehensive comparative study of machine learning algorithms implemented in the WEKA (Waikato Environment for Knowledge Analysis) platform for breast cancer prediction. The Breast Cancer Wisconsin (Diagnostic) dataset from the UCI Machine Learning Repository is used as the experimental dataset. Various classification algorithms — Logistic Regression, Naive Bayes, Decision Tree (J48), Random Forest, k-Nearest Neighbors (IBk), Support Vector Machine (SMO), and Multilayer Perceptron (MLP) — were evaluated based on accuracy, precision, recall, F1-score, and ROC area using 10-fold cross-validation. Results indicate that Random Forest and SMO achieved the highest prediction accuracy above 97%, demonstrating excellent generalization capability. The study concludes that WEKA provides an effective environment for estimating and comparing machine learning models in healthcare diagnostics.