Prediction of breast cancer survivability using ensemble algorithms
Vincent F. Adegoke, Daqing Chen, Ebad K. Banissi, Safia Barikzai · 2017
In this paper, several ensemble cancer survivability predictive models are presented and tested based on three variants of AdaBoost algorithm. In the models we used Random Forest, Radial Basis Function Network and Neural Network algorithms as base learners while AdaBoostM1, Real AdaBoost and MultiBoostAB were used as ensemble techniques and ten other classifiers as standalone models. There has been major research in ensemble modelling in statistics, medicine, technology and artificial intelligence in the last three decades. This might be because of the effectiveness and reliability of the technique in medical diagnosis and incident predictions compare with the standalone classifiers. We used Wisconsin breast cancer dataset in training and testing the models. The performances of the ensemble and standalone models were evaluated using Accuracy, RMSE and confusion matrix predictive parameters. The result shows that despite the complexity of the ensemble models and the required training time, the models did not outperform most of the standalone classifiers.