Caret Logistic model results in NULL model

Question

I am not sure what is happening here. When I create the model outside of caret it seems to work ok. However when using caret to get a cv I get NULL when I get a summary of the model. I also get an error when I try to plot the model. The response variable is a factor

data_ctrl = trainControl(method = "cv", number = 10)

model_caret1 = train(Clicked.on.Ad~ Age+ Area.Income+Daily.Internet.Usage + Daily.Time.Spent.on.Site,
                     data = Ads,
                     trControl = data_ctrl,
                     method = "glm",
                     family=binomial())

It results in the following:

all:
NULL

Deviance Residuals: 
    Min       1Q   Median       3Q      Max  
-2.4578  -0.1341  -0.0333   0.0167   3.1961  

Coefficients:
                            Estimate  Std. Error z value             Pr(>|z|)    
(Intercept)              27.12906491  2.71436398   9.995 < 0.0000000000000002 ***
Age                       0.17092126  0.02568321   6.655    0.000000000028334 ***
Area.Income              -0.00013539  0.00001868  -7.247    0.000000000000425 ***
Daily.Internet.Usage     -0.06391289  0.00674508  -9.475 < 0.0000000000000002 ***
Daily.Time.Spent.on.Site -0.19192952  0.02065752  -9.291 < 0.0000000000000002 ***
---
Signif. codes:  0 ‘***’ 0.001 ‘**’ 0.01 ‘*’ 0.05 ‘.’ 0.1 ‘ ’ 1

(Dispersion parameter for binomial family taken to be 1)

    Null deviance: 1386.3  on 999  degrees of freedom
Residual deviance:  182.9  on 995  degrees of freedom
AIC: 192.9

Number of Fisher Scoring iterations: 8

> plot(model_caret1)
Error in plot.train(model_caret1) : 
  There are no tuning parameters for this model.

The plot function will plot the corss-validation error as a function of the tuning parameters you've specified in the training. Since you haven't specified any tuning parameters, because the logistic regression model doesn't have any, you don't get any plot output. There is a discussion of this [here](https://stackoverflow.com/questions/47822694/logistic-regression-tuning-parameter-grid-in-r-caret-package). — DaveArmstrong, Dec 20 '20 at 16:17

score 1 · Answer 1 · answered Jan 04 '21 at 10:12

The object you obtained is a train object and i am guessing what you need to do is a plot on the glm object. So you need to look for the final fitted model:

library(caret)
dat = iris
dat$Species = factor(ifelse(dat$Species=="versicolor","v","o"))

data_ctrl = trainControl(method = "cv", number = 10)

model_caret1 = train(Species ~ .,
                     data = dat,
                     trControl = data_ctrl,
                     method = "glm",
                     family=binomial())

class(model_caret1)
[1] "train"         "train.formula"

Then under this:

class(model_caret1$finalModel)
[1] "glm" "lm" 

summary(model_caret1$finalModel)

plot(model_caret1$finalModel)

Caret Logistic model results in NULL model

1 Answers1