Parkinson’s Disease Detection Using MDVP and Different Types of ML Classification Algorithms

Parkinson’s Disease (PD) is a neurological degenerative disorder that is progressive in nature which primarily affects the motor system present in our body. The cause of spread of this disease is not contagious or infectious spread but is caused because of complex interactions between the human genetics with the surrounding environmental factors, leading to the disintegration of neurons present in our brain. Increase in number of patients of PD, caused researchers to implement use of various machine learning algorithms to detect and analyse PD using audio input and Magnetic Resonance Imaging (MRI)/(PET) or (DAT) scans. The main aim is a system designed and developed as a disease detection method for PD, and analyses the patient audio input and behavioural patterns. The proposed system uses SVC, RF and Various other classifier algorithms to build the classifier to detect the disease. To handle data and to ensure a good level of detection error and optimal training time, a pre-processing step and data analysis is used. Later this dataset is divided into train and test datasets. Our model gives the final accuracy of 94.87% using soft voting and have a F1 score percentage of 96.9%.