![]() | Only 14 pages are availabe for public view |
Abstract This study developed hybrid model (DT-SVM)using data mining techniques to predict the survival of advanced cancer patients. In order to give an early indication of whether advanced cancer patients might be surviving less than or equal to 30 days or more than 30 days, then predict the duration of survival per days within each class. We developed attribute importance model, the output used to develop decision tree model using C5.0 algorithm, with accuracy of 79 %, precision of 81% and recall of 70%. Then build the support vector machine model for each class. Both SVM models selected different 10 input variables represent the predictor importance, with accuracy of 74% and 73%. |