Diagnosing breast cancer using machine learning

Abstract

Breast cancer is a common global disease that contributes significantly to women's mortality due to misdiagnosis or delayed treatment, necessitating an urgent need for accurate diagnosis. Machine learning (ML) techniques are an important field of study proven effective in cancer prediction and early diagnosis. This study aimed to diagnose breast cancer using ML models and evaluate their effectiveness based on four criteria: accuracy, specificity, sensitivity, and F1 score. The research utilized a dataset from Kaggle, originally collected by Dr. William H. Wolberg at the University of Wisconsin Hospital in the United States. Four ML algorithms were tested: Random Forest (RF), Gradient Boosting Classifier (GB), Logistic Regression (LR), and Support Vector Machine (SVM). Upon comparing the results, the Random Forest algorithm was concluded to have achieved the highest accuracy, reaching 96.4%.