Author(s): Shorouq Eletter, Tahira Yasmin, Ghaleb Elrefae, Abdullah Elrefae
Healthcare organizations currently employ predictive analytics to improve patient care. Using data mining models to predict breast cancer is significant for improved care outcomes and patient experience.Providing data-driven decision-making process. Furthermore, historical records help to create models that aid disease detection at an early stage, which might also positively impact care outcome and patient experience. Breast cancer screening is a common service offered by healthcare providers. The study aims to evaluate the performance of five data mining algorithms for the prediction of breast cancer. Five data mining namely deep learning, naïve base, generalized linear models, support vector machines and random forest are applied to predict breast cancer. Four metrics will be used to assess the performance and highlight the best classification model. Deep learning yielded higher performance results on all metrics and ranked the variables based on their importance in building the classifier. Deep learning algorithms can be used successfully to predict breast cancer. Glucose and Resistin were the most important in the classification process. Therefore, high levels of these variable need to be monitored.