PREDICTING MILD COGNITIVE IMPAIRMENT AND DEMENTIA: MACHINE LEARNING VERSUS TRADITIONAL METHODOLOGIES
2024

Predicting Mild Cognitive Impairment and Dementia Using Machine Learning

Sample size: 4975 publication Evidence: moderate

Author Information

Author(s): Kang Bada, Hong Dahye, Kim Jennifer, Oh Sarah

Primary Institution: Yonsei University

Hypothesis

Can machine learning models predict mild cognitive impairment and dementia more accurately than traditional methodologies?

Conclusion

Machine learning models, especially XGBoost, show promise in predicting mild cognitive impairment and dementia onset more effectively than traditional methods.

Supporting Evidence

  • Logistic regression was effective in predicting dementia onset.
  • XGBoost outperformed other models in predicting mild cognitive impairment.
  • Increased assets were associated with reduced MCI risk.
  • Lower education and reduced daily living activities increased MCI risk.
  • Being female and older age were predictors of developing dementia.

Takeaway

This study found that using computers to analyze data can help doctors figure out if older people might have memory problems or dementia better than older methods.

Methodology

The study used the Korean Longitudinal Study of Aging dataset to compare machine learning models with logistic regression for predicting MCI and dementia.

Participant Demographics

Older adults aged 60 years or older from Korea.

Digital Object Identifier (DOI)

10.1093/geroni/igae098.3323

Want to read the original?

Access the complete publication on the publisher's website

View Original Publication