Performance Evaluation of Ensembles Algorithms in Prediction of Breast Cancer

Abstract

Breast Cancer is the most dominant cause of mortality in women. Early diagnosis and treatment of the disease can stop the spreading of cancer in the breast. Due to this nature of the problem, accurate prediction is the most important measure of the predictive model. This paper proposes the comparison of ensemble learning techniques in predicting breast cancer. This approach is widely used for performance improvement of the predictive task. The ensembles algorithms used in this research study are AdaBoost, Random Forest, and XGBoost. The result indicates that the random forest is the best predictive model for this dataset. The model has the following performance measure, accuracy 97%, sensitivity 96%, and specificity 96%. The experiment is executed using scikit-learn machine learning library. With this high level of accuracy offered by the model, the model can help the doctor to identify whether the patient has malignant or benign tumor cancer cells with high precision.