A hybrid classification model for predicting breast cancer survivability using public dataset

Abstract

Breast cancer has been one of the deadliest diseases and becomes the leading cause of cancer deaths in women worldwide. It becomes an issue to any countries as the number of victims developing breast cancer is increasing, and the survivability rate of cancerous women is still ambiguous. In this study, a hybrid classification model is developed for predicting breast cancer survivability. The hybrid classification model is a combination of three single classification models, namely Naive Bayes, J48 Decision Tree and RBF Network. Stacking method has been used to combine those models. This hybrid model has been tested using Breast Cancer Wisconsin dataset which gives result 96.14% of accuracy, 96.30% of precision, 96.10% of recall and 96.20% of f-measure. These results are outperformed each of the single classifiers.