Using Machine Learning to Analyze Voice Disorders for Parkinson's Disease Detection

A destructive neurodegenerative ailment that affects the nervous system and the body's nerve-controlled areas is known as Parkinson's disease. It is an autoimmune neurological condition that results in unintentional or uncontrolled motions, including rigidity, trembling, and trouble balancing and coordinating. As the population ages, managing Parkinson's disease (PD) is expected to become a more significant and difficult part of neurology and general practice for medicine specialists. PD is still difficult to diagnose, and research is still being done to characterize the condition in its early stages. Risk elements for developing Parkinson's include age, genetic background, exposure to pesticides, and being subjected to environmental toxins. PD patients frequently demonstrate a combination of non-motor as well as motor symptoms, including stiffness, seizure, bradykinesia, and stooped posture. Here, we intend to develop the use of a model trained with machine learning that, from speech recordings, can reliably determine whether a person has Parkinson's disease (PD) as the Decision tree achieved an accuracy of 94 %, Logistic Regression achieved 91%, KNN achieved 97% and Naive Bayes achieved 94%.