No Detail Found!

Classification models comparison for parkinson's disease detection through speech features

Author(s)
CHRISTOPHER PUTRA SETIAWAN Computer Science Department, BINUS Graduate Program - Master of Computer Science, Bina Nusantara University, Jakarta, Indonesia
EDWARD PRATAMA PUTRA Computer Science Department, BINUS Graduate Program - Master of Computer Science, Bina Nusantara University, Jakarta, Indonesia
GEDE PUTRA KUSUMA Computer Science Department, BINUS Graduate Program - Master of Computer Science, Bina Nusantara University, Jakarta, Indonesia
Abstract
Prior researches have used speech features to classify Parkinson's disease. This research was conducted to compare thirteen classification models in order to achieve a better performance in the newly published dataset. The method offered in this paper involves the usage of hyperparameter tuning, feature selection using genetic algorithm, and model stacking with ANN as the final classifier. The final stacked model has achieved an accuracy of 90.13% in the test dataset, with an F1- score of 93.7%. The results indicate that classification accuracy of Parkinson's disease classification through speech features can be increased by utilizing hyperparameter tuning, feature selection, and ensemble stacking. This improvement also indicated that the usage of feature selection in tandem with ensemble stacking method can yield better result compared to prior state of the art models for Parkinson's disease detection.
Publication Details
Page(s) 5386-5394
DOI DOI not available
Published Journal: Journal of Theoretical and Applied Information Technology, Volume: 100, Issue: 17, Year: 2022

Request the full text from PASTIC

Keywords
machine learning, Feature selection, Ensemble Stacking, Parkinson's Disease Detection, Hyperparameter Tuning
More from Journal of Theoretical and Applied Information Technology
0

Citations

0

Downloads

40

Views

Copyright ©  PASTIC National Center, Islamabad www.pastic.gov.pk