Skip to content

Parkinson diagnostic with supervised and unsupervised machine learning

Notifications You must be signed in to change notification settings

MaryemSamet/Parkinson-diagnostic

Repository files navigation

Parkinson-diagnostic

Parkinson’s Disease diagnosis classification using gait cycle similarities

The objectif is to do Feature Extraction from the orginal dataset.

Then I used the new generated dataset with the new features with different classification machine learning algorithms to predict parkinson's disease.

I used Similarity measures like 'Euclidean Distance' ,'Time-Wraped Edit Distance', 'Dynamic Time Wrapping', 'Edit Distance on Real sequence': to measure similarity between Timeseries Data.

The timeseries data represent the Force VGRF(Vertical Ground Reaction Force) of people while walking.

*Feature extraction with Similarity Measurement (on Timeseries data)

*Dealing with unbalanced data

*Using Unsupervised and supervised Learning

*evaluate models

Dataset source : https://physionet.org/content/gaitpdb/1.0.0/