FEATURE SELECTION USING EXTRA TREES CLASSIFIER FOR PARKINSON’S DISEASE CLASSIFICATION

JOURNAL OF MECHANICS OF CONTINUA AND MATHEMATICAL SCIENCES, Journal Year: 2024, Volume and Issue: spl11(1)

Published: May 24, 2024

Parkinson's disease (PD) is chronic, permanent, and life-threatening. Neurologically protective treatments for PD rely on early detection. Recent studies have demonstrated that clinical data, cerebrospinal Fluid (CSF) based proteomes, gene mutations are important biomarkers accurate detection of PD. This study aims to investigate the heterogeneous data comprised CSF-based proteomic analysis as well mutation information genes, Glucose Beta Acid (GBA), leucine-rich kinase (LRRK2) classify controls into PD-affected Healthy Control (HC). The dataset contains 1103 (569 affected 534 HC). Automated Machine Learning (AutoML) framework using PyCaret utilized. has proposed an Extra Tree Classifier (ETC) a feature selection mechanism select features significantly affect classification. Selected further used train Random Forest (RF), Logistic Regression (LR), Decision (DT) classifiers. Accuracy, sensitivity, specificity, area under receiver operating characteristic curve (AUC-ROC), confusion matrix evaluate performance RF depicted best in terms accuracy value 96.12%, sensitivity 95.59%, specificity 95.34% while LR shown highest AUC 98.33. made number correct predictions 316 out 331.

Language: Английский

Enhanced Parkinson’s Disease Diagnosis via MRI Analysis: Integrating Deep Features From DenseNet201 With Neural Network Techniques

Applied Computational Intelligence and Soft Computing, Journal Year: 2024, Volume and Issue: 2024(1)

Published: Jan. 1, 2024

Parkinson’s disease (PD) is a neurodegenerative disorder that affects millions of people worldwide, necessitating accurate and timely diagnostic methods for effective management. This study proposes novel approach PD detection using deep features extracted from magnetic resonance imaging (MRI) scans, employing pattern recognition neural network architecture based on DenseNet201. The developed model demonstrated exceptional performance, achieving validation test accuracies 99.4% 99.2%, respectively, indicating its robustness efficacy in distinguishing between patients with healthy individuals. Furthermore, the achieved precision 99.5%, recall 99.3%, an F1 score 99.4%. For set, accuracy was at 99.1%, 99.2%. rapid convergence during training further underscores efficiency learning discriminative MRI images. These findings underscore promising role techniques, particularly convolutional networks (CNNs), medical image analysis diagnosis. proposed holds significant potential assisting clinicians early diagnosis personalized treatment planning, ultimately improving patient outcomes quality life.

Language: Английский

Exploring the Potential Imaging Biomarkers for Parkinson’s Disease Using Machine Learning Approach

Bioengineering, Journal Year: 2024, Volume and Issue: 12(1), P. 11 - 11

Published: Dec. 27, 2024

Parkinson's disease (PD) is a neurodegenerative disorder characterized by motor and neuropsychiatric symptoms resulting from the loss of dopamine-producing neurons in substantia nigra pars compacta (SNc). Dopamine transporter scan (DATSCAN), based on single-photon emission computed tomography (SPECT), commonly used to evaluate dopaminergic striatum. This study aims identify biomarker DATSCAN images develop machine learning (ML) algorithm for PD diagnosis. Using 13 DATSCAN-derived parameters patient handedness 1309 individuals Progression Markers Initiative (PPMI) database, we trained an AdaBoost classifier, achieving accuracy 98.88% area under receiver operating characteristic (ROC) curve 99.81%. To ensure interpretability, applied local interpretable model-agnostic explainer (LIME), identifying contralateral putamen SBR as most predictive feature distinguishing healthy controls. By focusing single biomarker, our approach simplifies diagnosis, integrates seamlessly into clinical workflows, provides interpretable, actionable insights. Although has limitations detecting early-stage PD, demonstrates potential ML enhance diagnostic precision, contributing improved decision-making outcomes.

Language: Английский