Parkinson's Disease is a nervous system disorder and is progressive in nature. It affects bodymovements. Symptoms begin with a hardly noticeable tremor in body parts. The disorder resultsin stiffness and slowing down of movements. It’s a debilitating neurodegenerative disease and cannot be diagnosed through blood tests. Parkinson’s disease mostly affects the people above 60 and is one of the common diseases among war veterans. Hence, there was a need for a faster and cheaper diagnostic tool. The project uses ML algorithms to analyze the variations in voice patterns to predict the existence of Parkinson's Disease in the patients. Pearson method of correlation is used to find out the best features and an ensemble model(XGBoost) is used to diagnose Parkinson's Disease with maximum accuracy using a dataset that consists of data from voice recordings of Parkinson's patients and unaffected subjects. These data of varying frequencies can be fed to the model and the results can be compared to find the people who are affected with the PD and display the result.

Keywords—Parkinson’s Disease, voice patterns, XGBoost.