Breast cancer prediction based on neural networks and extra tree classifier using feature ensemble learning

Abstract

Cancer prediction has always been a major and difficult matter for doctors and researchers. Early-stage detection of disease can help in the timely diagnosis and prognosis. Several methods for cancer's early prediction have been proposed by various researchers. In this paper, the author proposed a feature called ensemble learning based on neural networks and extra trees for the classification of breast cancer into non-cancerous (i.e. benign) and cancerous (i.e. malignant). Breast Cancer Wisconsin (Diagnostic) medical data sets from the machine learning repository have been used. The performance of the proposed method is evaluated with indices like accuracy in classification, specificity, sensitivity, recall, precision, f-measure, and MCC. Simulation and results have established that the accepted approach is more capable of giving results when different parameters are selected. The prediction results obtained by the proposed approach were very propitious (99.74% true accuracy). In addition, the proposed method, Neural Network and Extra Tree (NN-ET) outperforms other state-of-the-art classifiers in terms of various performance indices. The model suggested has proved to be more efficient and beneficial for breast cancer classification, which is also shown by experimental simulations, empirical results, and statistical analyses. It is also compared to the machine learning models that are already out there in the related literature.