ppmi-preprocess
The project is about how to preprocess the clinical data of Parkinson's Progression Markers Initiative (PPMI). The tables can be categorized as demographics, motor, non-motor, biospecimen, neuroimaging, and medication features. Due to the required permissions of the Michael J. Fox Foundation (MJFF), we are not going to provide the data. Our aim is to make the preprocess easier.
The project currently includes file reading, feature concatenation, and imputation. The generated ".pkl" files includes sequential features for each patient and her/his associated labels (PD or control, Hoehn & Yahr stage, or cognitive impairment scale). The only step you need to do is following the "filename" lists under the subfolders, copying the corresponding raw ".csv" files into the folder "data", and then run main.py.