New Survival Prediction System for Terminal Patients based on Machine Learning

In recent years, because cancer treatment and palliative care have been conducted together, suitable and long-term survival prediction is required for cancer treatment and palliative care planning. Until now, the Palliative Prognostic Index (PPI) has been used for survival prediction systems. Therefore, in this study, we extracted prognostic factors using machine learning and Random Forest regression and created survival prediction systems using new factors (i.e., based on those extracted) that cannot be used for long-term prediction in PPI.Results show that the new prediction systems yielded higher accuracies thanPPI.