| Summary: | This paper present ADHD detection using a machine learning algorithm. ADHD is a prevalent condition affecting many children and adults globally, with the rate of undiagnosed persons increasing tremendously. The prime method to diagnose and confirm ADHD is still clinically driven, requiring specialists in short supply, with the diagnostic process taking months to complete. We utilized a machine learning (ML) algorithm to classify or detect ADHD cases using recorded activity data in the HYPERAKTIV dataset. The selected ML models have been shown to perform at least comparable to the prior studies, with 82% or higher accuracy than the data provider's 72% accuracy. Additional data processing and augmentation have been demonstrated to increase the performance of the models for a few algorithms. The combination of findings in the paper is hoped to path a way to provide closure to ADHD persons by providing initial classification provided the similar data format based on activity data becomes more accessible through technological advancements such as smartphones and wearable devices. © 2022 IEEE.
|
|---|