Stack Overflow en español Asked by la_roca on October 2, 2020
Queria preguntarles como hacer un método de validacion cruzada simplificado, con solo datos de entrenamiento y testeo, especificamente seleccionando 100.000 variables seleccionadas aleatoriamente como entrenamiento, y el resto para testeo e implementar la validacion cruzada 200 veces. Esto no es separando posteriormente los datos de entrenamiento, nuevamente en entrenamiento y validación?
Agradezco su ayuda!
Para realizar la validación cruzada, sólo se hace con el grupo de entrenamiento y no con el testeo.
library(caret)
set.seed(2020)
index <- createDataPartition(datos$target, p=0.7, list=FALSE)
train<- datos[index,]
test<- datos[-index,]
Con esto se logra dividir la base general en entrenamiento y testeo.
Ahora para poder realizar la validación cruzada sería de la siguiente manera:
fit_control <- trainControl(method = "cv", number = 200, savePredictions = 'final', classProbs = T)
Con esto realizas una validación cruzada 200 veces.
Answered by césar huamani ninahuanca on October 2, 2020
1 Asked on November 10, 2021
1 Asked on November 10, 2021 by leen
1 Asked on November 10, 2021 by chrlord
2 Asked on November 10, 2021
1 Asked on November 7, 2021 by peter-paul
1 Asked on November 7, 2021 by fernando-de-jesus-basurto
1 Asked on November 7, 2021
2 Asked on November 7, 2021 by maha1982
1 Asked on November 7, 2021 by miquel
1 Asked on November 7, 2021 by alejandro-erazo-bolaos
1 Asked on November 7, 2021 by hoover
0 Asked on November 7, 2021 by tupi
0 Asked on November 7, 2021 by jorge-gonzales
1 Asked on November 7, 2021
1 Asked on November 7, 2021 by carlos-daniel-zrate-ramrez
1 Asked on November 7, 2021 by kenny-cifuentes
2 Asked on November 7, 2021 by nassif-113
0 Asked on November 7, 2021 by ivan-manuel-morillo
1 Asked on November 4, 2021 by erain-moya
Get help from others!
Recent Answers
© 2022 AnswerBun.com. All rights reserved. Sites we Love: PCI Database, MenuIva, UKBizDB, Menu Kuliner, Sharing RPP