Résumé
In this study, we combined a Poisson regression model with neural networks (neural network Poisson regression) to relax the traditional Poisson regression assumption of linearity of the Poisson mean as a function of covariates, while including it as a special case. In four simulated examples, we found that the neural network Poisson regression improved the performance of simple Poisson regression if the Poisson mean was nonlinearly related to covariates.We also illustrated the performance of the model in predicting five-year changes in cognitive scores, in association with age and education level; we found that the proposed approach had superior accuracy to conventional linear Poisson regression. As the interpretability of the neural networks is often difficult, its combination with conventional and more readily interpretable approaches under the generalized linear model can benefit applications in biomedicine.
| Langue d'origine | English |
|---|---|
| Pages (de-à) | 2051-2062 |
| Nombre de pages | 12 |
| Journal | Journal of Applied Statistics |
| Volume | 38 |
| Numéro de publication | 9 |
| DOI | |
| Statut de publication | Published - sept. 2011 |
Note bibliographique
Funding Information:Authors wish to thank Ruth Ripley and Ian Nabney for their help for R and Matlab Codes. Authors wish to thank Krisa Patriquin for help with editing. These analyses were supported by the Canadian Institutes for Health Research (CIHR) operating grant MOP-64169. Additional support for Nader Fallah came from a postdoctoral fellowship of Alzheimer’s Society of Canada and a Mathematics of Information Technology and Complex Systems (MITACS) postdoctoral award. The decision to submit this paper was entirely the decision of the authors and not the sponsors, who had no role in design, analysis or writing of the paper.
ASJC Scopus Subject Areas
- Statistics and Probability
- Statistics, Probability and Uncertainty