Intervention in prediction measure: a new approach to assessing variable importance for random forests

BACKGROUND: Random forests are a popular method in many fields since they can be successfully applied to complex data, with a small sample size, complex interactions and correlations, mixed type predictors, etc. Furthermore, they provide variable importance measures that aid qualitative interpretati...

Descripción completa

Detalles Bibliográficos
Autor principal: Epifanio, Irene
Formato: Online Artículo Texto
Lenguaje:English
Publicado: BioMed Central 2017
Materias:
Methodology Article
Acceso en línea:https://www.ncbi.nlm.nih.gov/pmc/articles/PMC5414143/
https://www.ncbi.nlm.nih.gov/pubmed/28464827
http://dx.doi.org/10.1186/s12859-017-1650-8

Internet

https://www.ncbi.nlm.nih.gov/pmc/articles/PMC5414143/
https://www.ncbi.nlm.nih.gov/pubmed/28464827
http://dx.doi.org/10.1186/s12859-017-1650-8

Ejemplares similares