AdaBoost for Parkinson's Disease Detection using Robust Scaler and SFS from Acoustic Features

Parkinson's disease is a degenerative neurological state that begins with body part movements. The brain controls dopamine, which allows for seamless and synchronous action of the muscular system. Dopamine is produced in the brain's “Neuron loss cells,” which are one of the brain's components. When a person has Parkinson's disease, those cells begin to die. Dopamine pathways will be reduced in the future. When the proportion is reduced from 60 to 80, symptoms begin to occur. The suggested method focuses on standardizing the given data, sequential feature selector analyzing selected characteristics. The number of features is reduced to ten, and other classifier algorithms are implemented, including Random Forest, Logistic Regression, SVM, Gaussian Naive Bayes and KNN. The Robust Scaler scales the features by robust outliers. The SFS feature selection technique, Randomized Search CV, and AdaBoost classifier produced best results, with a 100% accuracy rate. The Adaboost classifier achieves an AUC-ROC score of 1.0, as well Gradient boosting also achieves a score of 1.0, which is our biggest win when utilizing ten acoustic features.