ZarishWajid, BilalRashid, UmarZahid, SajidaAnwar, FariaAwan, Fahim GoharAnwar, Abdul RaufWajid, Imran2023-10-272023-10-272022https://hdl.handle.net/11467/6840https://doi.org/10.1109/ETECTE55893.2022.10007402Survival rate prediction for medical diseases is a complex task that requires high precision. With a low survival rate among reported patients, leukemia is a type of cancer of blood which is caused by the abnormal growth of white blood cells. It is critical to numerically evaluate the rate of survivability of patients suffering from leukemia. To this end, this paper employs a comprehensive database, namely Surveillance, Epidemiology, and End Results (SEER) maintained by The National Cancer Institute in MD, USA, to construct a survivability model for leukemia patients. To accurately predict the survival months of the patients, we develop a multi-class classification problem by binning the target variable into four bins. The resulting accuracy is improved by utilizing a multi-tier classification framework. Although, the final numerical results hold significance from biological viewpoint, it is recommended that a clinically relevant model be drawn with caution.eninfo:eu-repo/semantics/embargoedAccessBlood cancer, leukemia, lymphoma, machine learning, survival months, SEERConference ObjectN/A2-s2.0-8514714045710.1109/ETECTE55893.2022.10007402