Performance evaluation of Dictionary Learning and ICA on Parkinson’s patients classification using Machine Learning

• Published in: Multimedia tools and applications Vol. 83; no. 8; pp. 24467 - 24483
• Main Authors: Dutta, Saloni Bhatia, Vig, Rekha
• Format: Journal Article
• Language: English
• Published: New York Springer US 01.03.2024; Springer Nature B.V
• Subjects: Algorithms; Classification; Computer Communication Networks; Computer Science; Data Structures and Information Theory; Dictionaries; Feature extraction; Independent component analysis; Machine learning; Magnetic resonance imaging; Multimedia Information Systems; Parkinson's disease; Performance evaluation; Principal components analysis; Special Purpose and Application-Based Systems; Track 2: Medical Applications of Multimedia; Feature Extraction; CanICA; Dictionary Learning; Functional Connectivity; Machine Learning
• ISSN: 1573-7721; 1380-7501; 1573-7721
• DOI: 10.1007/s11042-023-16485-5

Currently, extensive research is being conducted in the application of Machine Learning (ML) algorithms in the medical domain and one such area is classifying Parkinson’s Patients. Feature extraction methods play a critical part in the use of ML techniques in providing better accuracies. The earlier studies have used various feature extraction techniques like Principal Component Analysis (PCA), Independent Component Analysis (ICA), Region of Interest (ROIs), etc. This paper deals with the use of two feature extraction techniques - Canonical Independent Component Analysis (CanICA) and Dictionary Learning (DL) for the Functional Magnetic Resonance Imaging (fMRI) modality. Region of Interest (ROI) extraction of connected components and dimensionality reduction algorithms further refine the features. The features extracted are then applied to the Machine Learning models for the classification of individuals suffering from Parkinson’s disease. The methodology adopted in the research provided accuracy of 87.5% and 86.6% using the CanICA and DL techniques respectively. The accuracies obtained are found to be better than the other research conducted using ML algorithms for the MRI data.