Comparison of neural network and logistic regression for dementia prediction: Results from the preadvise trial

Research output: Contribution to journalArticlepeer-review


Objective. Two systematic reviews suggest that current parametric predictive models are not recommended for use in population dementia diagnostic screening. This study was to compare predictive performance between logistic regression (conventional method) and neural network (non-conventional method). Method. Neural network analysis was performed through the R package “Neuralnet” by using the same covariates as the logistic regression model. Results. Results show that neural network had a slightly apparently better predictive performance (area under curve (AUC): 0.732 neural network vs. 0.725 logistic regression). Neural network performed similarly as logistic regression. Furthermore, logistic regression confirmed that the interaction effect among covariates, which elucidated from neural network. Conclusions. Neural network performed slightly apparently better than logistic regression, and it is able to elucidate complicated relationships among covariates.

Original languageEnglish
Pages (from-to)137-146
Number of pages10
JournalJournal of Gerontology and Geriatrics
Issue number2
StatePublished - Jun 2021

Bibliographical note

Funding Information:
We thank Dr. Erin Abner for her assistance with helpful comments. Financial Disclosure: PREADViSE (NCT00040378) is supported by National Institute on Aging (NIA) R01 AG019421. Additional support for the current study comes from NIA R01 AG038651 and P30 AG028383. SELECT was supported by NCI grants CA37429 and UM1 CA182883. NCI was involved in the design of SELECT.

Publisher Copyright:
© by Società Italiana di Gerontologia e Geriatria (SIGG).


  • Dementia
  • Logistic regression
  • Neural network
  • Prediction

ASJC Scopus subject areas

  • Aging
  • Geriatrics and Gerontology


Dive into the research topics of 'Comparison of neural network and logistic regression for dementia prediction: Results from the preadvise trial'. Together they form a unique fingerprint.

Cite this