الفهرس | Only 14 pages are availabe for public view |
Abstract One of the most prevalent forms of cancer among women is breast cancer. Early and precise detection can minimize the impact on the health of patients. Therefore, Machine learning approaches can substantially improve the process of early cancer diagnosis and prediction. This study focuses on the use of machine learning techniques for the prediction and detection of breast cancer. The proposed model involves applying a set of nine distinct ML based classification models such as Logistic Regression, Support Vector Machine, Decision Tree, Random Forest, Naïve Bayes, Extreme Gradient Boosting, Adaptive Boosting and Artificial Neural Network. Several experiments were conducted in this study including different data splitting sizes and feature selection methods. The assessment of these models has been done based on four performance metrics including accuracy, precision, f-measure, and recall. The first proposed model for binary classification of breast cancer indicated that XGB yielded the highest scores of 99.73% in terms of accuracy while using a fewer number of predictive features. Moreover, Logistic Regression achieved an accuracy of 98.24% without implementation of feature selection technique. While the neural network reached the highest accuracy of 98.68% and outperformed the remaining techniques. The second proposed model for multi classification of breast cancer achieved an accuracy of 93.95% when using AdaBoost with feature selection. |