Multiple Sclerosis Prediction of the Proposed Hybrid GRUCB Method Pooling with Catboost and Gated Recurrent Unit

Abstract

Multiple Sclerosis (MS) is a persistent immunological disorder that affects the central nervous system (CNS) which causes damages to the nervous system such as brain and Spinal Cord. Immune system is protected by protective layer known as myelin sheath. When this myelin sheath is affected by some infectious agent then it results MS. MS causes inflammation and damages that disrupting the transmission that link the brain and the remaining body parts. There are several reasons for MS, some of them are due to genetic factors i.e., when certain gene related to the immune system (eg: HLA-DRB1) may get infected or by due to ecological conditionals such as Vitamin D inadequacy, geographical locations, Smoking and viral infections. Symptoms of MS includes Muscle Weakness, Fatigue, Numbness or Tingling, Bladder and Bowel Dysfunction, Vision Problems, Difficulty in walking and Cognitive Issues. There are several traditional diagnosis methods. Some of them includes Lumbar Puncture (Spinal Tap), Magnetic Resonance Imaging (MRI), Blood Tests and Evoked Potential Tests. As MS is a serious disorder which is growing fast, an alternative fast prediction method has been introduced with the help of ML and Deep Learning algorithm the prediction rate has been enhanced. By using proposed algorithm as the result, the accuracy rate achieved is 98%.