Please use this identifier to cite or link to this item:
|Determination of the Optimal EEG-based Features to Detect ADHD by Machine Learning Algorithms
|attention deficit hyperactivity disorder
Nearest neighbor search
Attention deficit hyperactivity disorder
Least absolute shrinkage and selection operators
Machine learning algorithms
Support vectors machine
Support vector machines
|Institute of Electrical and Electronics Engineers Inc.
|This study proposes a highly accurate and fast algorithm for the diagnosis of attention deficit hyperactivity disorder (ADHD), which will reduce reliance on time-consuming subjective assessments, the findings of which are likely to be mistaken with other neurodevelopmental diseases. Time, frequency and nonlinear features were extracted from electroencephalographic (EEG) signals which recording based on visual attention task obtained from 61 ADHD and 60 healthy participants. In this study, Least Absolute Shrinkage and Selection Operator (LASSO) was used to find reliable features; and four machine learning classifiers such as support vector machine (SVM), k-nearest neighbors (KNN), decision tree and ensemble learning were evaluated for classifying ADHD and healthy children. The results were indicated that using LASSO with SVM can be useful for classifying ADHD and the highest average accuracy was reached in this study was 96.3%. In addition, the features selected with LASSO had shown that signals from the temporal, parietal, and occipital lobes might have the possible biomarkers for ADHD, at least in tasks that require visual attention. © 2023 IEEE.
|2023 Medical Technologies Congress, TIPTEKNO 2023 -- 10 November 2023 through 12 November 2023 -- 195703
|Appears in Collections:
|Scopus İndeksli Yayınlar Koleksiyonu / Scopus Indexed Publications Collection
Show full item record
checked on Feb 26, 2024
Items in GCRIS Repository are protected by copyright, with all rights reserved, unless otherwise indicated.