Multiclass Classification of Glioblastoma and Medulloblastoma Tumors Using Machine Learning

Glioblastoma Multiforme and Medulloblastoma are malignant gliomas and are known as the most aggressive and challenging brain tumors to treat. The prognosis and treatment approach depends on various factors, including the type of glioma, its location, and the overall health of the patient. Although genetic alterations are present in all tumor types, the genes impacted and the processes leading to malignant transformation may vary. Multiple genetic and epigenetic events and gene dysregulation characterize tumor progression. This article aims to explore the significant genes responsible for both glioblastoma and medulloblastoma tumors by taking the approach of machine learning. Support Vector Machines algorithm was used to predict 4 such relevant genes, the correctness of which was further confirmed by existing literature and GEO2R analysis.