[R] How to extract parameters from train in caret package

Michela Leone mickyleo at gmail.com
Wed Nov 23 12:54:12 CET 2016


 Dear R-users,
   I need to extract coefficients  from the  final model after running
train function of caret package.

I need to have something like "summary" used after running a regression
model
in this case  from the best model.

below my  script   as explanation

data(Sonar)
set.seed(107)
 inTrain <- createDataPartition(y = Sonar$Class,  p = .75,  list = FALSE)

training <- Sonar[ inTrain,]
testing <- Sonar[-inTrain,]


 plsFit <- train(Class ~ .,
                   data = training,
                  method = "pls",
                  tuneLength = 15,
                  preProc = c("center", "scale"))

What I need is a summary  which looks like


Call:
glm(formula = Class ~ ., family = "binomial", data = training)

Deviance Residuals:
       Min          1Q      Median          3Q         Max
-2.418e-05  -2.110e-08  -2.110e-08   2.110e-08   2.697e-05

Coefficients:
              Estimate Std. Error z value Pr(>|z|)
(Intercept)  1.081e+02  2.668e+05   0.000    1.000
V1           7.235e+02  4.390e+06   0.000    1.000
V2          -1.024e+03  2.282e+06   0.000    1.000
V3           1.113e+03  2.735e+06   0.000    1.000
V4          -7.072e+02  1.828e+06   0.000    1.000
V5          -1.542e+01  1.534e+06   0.000    1.000
V6          -1.983e+02  1.438e+06   0.000    1.000
V7           2.543e+02  1.465e+06   0.000    1.000
V8           3.507e+02  8.199e+05   0.000    1.000
V9          -4.010e+02  8.517e+05   0.000    1.000
V10          4.317e+01  9.660e+05   0.000    1.000
V11          1.364e+02  1.800e+06   0.000    1.000
V12         -4.606e+02  2.350e+06   0.000    1.000
V13          1.746e+02  1.246e+06   0.000    1.000


    Null deviance: 2.1688e+02  on 156  degrees of freedom
Residual deviance: 1.0630e-08  on  96  degrees of freedom
AIC: 122

Number of Fisher Scoring iterations: 25



Many thanks for your help ! Michela Leone

	[[alternative HTML version deleted]]



More information about the R-help mailing list