Using Machine Learning to Identify Common Engagement-Related Behaviours Demonstrated by Older Adults with Dementia While Playing Mobile Games

• Author / Creator

  Torabgar, Melika

• Background: Dementia causes impairment of a person’s memory, cognitive abilities, and behaviour, making it difficult for a person to complete daily tasks. Dementia affects the behavioural, psychological, and social dimensions of older adults living with the disease. Older adults living with dementia may demonstrate engagement through behaviours that differ from those without dementia when they participate in activities such as playing mobile games. This study aims to identify the most common engagement-related behaviours along with their personal characteristics, technical issues, and environmental disturbances to determine their applicability to dementia. This project could be a useful option to help rehabilitation professionals identify clients experiencing dementia based on their engagement-related behaviours while performing leisure activities such as mobile games.
  Methods: Participants included five individuals living with dementia and 10 individuals without dementia. The secondary analysis of a single case design was conducted. The Chi-squared, and Fisher’s exact tests were used for statistical analyses. Then, four Random Forest models were trained to identify the relevant engagement-related behaviours demonstrated by older adults with dementia from those without dementia. Random Forest Gini index was used to identify the strongest predictors of engagement-related behaviours of dementia.
  Results: 30/47 (64%) of engagement-related behaviours were statistically significantly different in the two groups (older adults living with/without dementia). The accuracy (F1 score) of the Random Forest models for identifying engagement-related behaviours demonstrated by older adults with dementia from those without dementia was 78% using engagement-related behaviours only, 88% using engagement-related behaviours along with personal characteristics, 79% using engagement-related behaviours along with environmental disturbances, and technical issues, and 91% using engagement-related behaviours, personal characteristics, technical issues, and environmental disturbances features. The area under the receiver operating curve for the final model was 99%.
  Conclusion: The findings show differences in frequencies of engagement-related behaviours demonstrated by older adults with and without dementia. The Random Forest model could be an accessible way to identify engagement-related behaviours commonly demonstrated by older adults with dementia while playing mobile games.

• Subjects / Keywords

  • Machine learning
  • Older adults
  • Engagement-related behaviours
  • dementia
  • mobile games

• Graduation date

  Fall 2022

• Type of Item

  Thesis

• Degree

  Master of Science

• DOI

  https://doi.org/10.7939/r3-danv-ez72

• License

  This thesis is made available by the University of Alberta Library with permission of the copyright owner solely for non-commercial purposes. This thesis, or any portion thereof, may not otherwise be copied or reproduced without the written consent of the copyright owner, except to the extent permitted by Canadian copyright law.