Parkinson’s Disease Prediction

Abstract

Voice-based biomarkers can help diagnose symptoms of dementia such as Parkinson's disease, PD is a modern neurodegenerative disease affecting about 7 million people worldwide (usually adults), with about 150 thousand new scientific diagnoses performed each year. Historically, PD has been difficult to find and documents tend to focus on a few symptoms and even ignore some, depending on the scores of independent points. Due to the decline in motor manipulation which is a sign of illness, the term can be used as a means of detecting and diagnosing PD. Common sense has meant that physicians often focus on the symptoms of PD while ignoring the other. By using independent measurement scales, the term can be used to diagnose and diagnose the disease. This paper presents evidence to support the concept of supervised classification, which can be used to diagnose individuals with diseases such as diabetes and pulmonary fibrosis. Through Linear Regression, Logistic Regression, Decision Trees, Support Vector Machine, Random Forest, XGBoost, Neural Network and Adaboost we were able to achieve a peak accuracy of 100% for diagnosing pathological conditions. The project also uses various Evaluation Methods and Metrics such as Confusion Matrix, Classification Report, F1 - Score, Accuracy, Precision, Recall.