About the Parkinsons Disease Stage Prediction Model

This model was trained on a limited set of ~ 37 participants across four locations within Long Island clinical sites.

ML approach: MlJar

MLJar is a Python package that offers a simple, easy-to-use interface for conducting machine learning experiments. It provides a number of popular algorithms for supervised and unsupervised learning, as well as tools for preprocessing and feature engineering. What makes MLJar particularly useful for simple, brief experiments is its emphasis on automation and ease of use. Users can quickly and easily train and evaluate models on their data, without needing to spend time on manual feature selection or hyperparameter tuning. MLJar also provides tools for model interpretation and visualization, which can help users gain insights into their data and models. Overall, MLJar is a great tool for those who want to conduct quick and simple experiments without sacrificing quality or accuracy.

Confusion Matrix (normalized)

Confusion Matrix

Learning Curves

Permutation Importance

Precision Recall Curve

ROC Curve

The model was trained on the following features:

#datacollectionlocation

#age

#gender

#heightininches

#weightinpounds

#race

#monthssincediagnosis

#handdominance

#timesincelastdoseofdopaminergicmeds

#numberoffallsinthelast6months

#numberoffallsinthelast3months

#12orfewyearsofeducation

#visuospatialexecutive

#attention

#language

#abstraction

#delayedrecall

#orientation

#mocatotal

#abctotalscore

#abccalculatedconfidence

#godinleisurescore

#f8wttrial1time

#f8wttrial2time

#f8wttrial3time

#f8averagetime

#f8trial1steps

#f8trial2steps

#f8trial3steps

#f8averagesteps

#10mtrial1time

#10mtrial2time

#10mtrial3time

#10maverage

#gaitspeedfrom10mwt

#complete

#falls_total

#falls_any

#mocatotal_recode

#abccaclulatedconfidence_recode

#godinleisurescore_recode

#gaitspeedfrom10mwt_recode

#missing_values