Diagnosis of Parkinson Disease using the spiral drawing test

There is interest in the use of simple bedside testing for the diagnosis of Parkinson disease. Several groups have applied machine learning methods using curated dataset to classify patients as having Parkinson disease or not. This project has several learning points. In this project, the student will learn to pull together multiple open sources dataset and evaluate the appropriateness of the dataset for classification task. For example, does the dataset have only Parkinson disease patients and healthy control or other type of patients who can serve as control. This is an important point as separating patients with Parkinson disease from these healthy controls can be easy but separating Parkinson disease from healthy control and disorders related to Parkinson disease may not be as easy. After compiling a dataset from the multiple open source dataset, the student will then apply machine learning method to this classification task. Machine learning can be performed using R or Python.