A More Effective Ensemble ML Method for Detecting Breast Cancer

Breast cancer is one of the more widespread carcinomas that affect women. The majority of afflicted women are above 50. Breast cancer was poised to become the world’s most common trigger of death by 2020, having been identified among 7.8 million women over the 5 years preceding that. ML algorithms are being used extensively in the medical field to forecast certain diseases sooner. A breast cancer dataset obtained from Kaggle is used in this study to test the effectiveness of eight different ML algorithms, including LR (Logistic Regression), GBC (Gradient Boosting Classifier), KNN (Knearest Neighbors Classifier), SVM (Support Vector Machine), RFC (Random Forest Classifier), ETC (Extra Tree Classifier), SC (Stacking Classifier), and VC (Voting Classifier). The most significant contribution of this research in this area is the creation of the stacking classifier (SC), an effective ensemble method that outperforms all other applicable techniques with an accuracy of 99.4152%. Another important contribution is hyper-parameter tuning which selects the best parameters for the above algorithms. Finally, the performance of all applying algorithms for this dataset is compared.