Predicting Sleep-Disordered Breathing Using Deep Learning Algorithms

Sara Jasen*, Evisi Nastasi, Sara Ghanim

*Corresponding author for this work

Research output: Contribution to book/anthology/report/proceedingBook chapterResearchpeer-review

1 Citation (Scopus)

Abstract

This paper aims to propose a prediction model for Sleep Disordered Breathing (SDB) in relation to different malocclusions and oral habits with the purpose of aiding clinicians in early screening of patients at risk of developing SDB. The study involved 398 children, 201 girls and 197 boys, aged 3 to 6 years old. Clinical examinations were conducted between February 2021 and November 2022 and included diagnosis of different malocclusions and oral habits. PSQ (Chervin’s Pediatric Questionnaire) was completed by the parents to assess SDB. Artificial Intelligence (AI) based Convolutional Feed Forward Neural Network (NN) SDB prediction model was created. Utilizing Python deep learning algorithms, a Multi-Layered Perceptron (MLP) was developed for SDB prediction. Out of 27 variables, 8 SDB predictors were identified. According to the NN, SDB predictors are: Angle Class II malocclusion, lower lip biting habit, lateral crossbite without mandibular shift, thumb sucking, male gender, increased overjet, open bite and atypical swallowing. The prediction model has an accuracy rate of 96.1%.

Original languageEnglish
Title of host publicationCutting-Edge Business Technologies in the Big Data Era : Proceedings of the 18th SICB “Sustainability and Cutting-Edge Business Technologies” Volume 2
Number of pages9
Place of publicationCham
PublisherSpringer
Publication date2023
Pages1-9
ISBN (Print)978-3-031-42457-1
ISBN (Electronic)978-3-031-42455-7
DOIs
Publication statusPublished - 2023
SeriesStudies in Big Data
Volume136
ISSN2197-6503

Keywords

  • artificial intelligence
  • deep learning
  • orthodontics
  • sleep disordered breathing

Fingerprint

Dive into the research topics of 'Predicting Sleep-Disordered Breathing Using Deep Learning Algorithms'. Together they form a unique fingerprint.

Cite this