Removing Insignificant factor and ordered factor variables from Logistic Regression Model

[karansehgal1988]

Hello All,

I have a query regarding the removal of insignificant factor variable and Ordered factor variable from regression model using R.

For Example -

1.) Normal regression model

a) Running the model using training data and I get the below summary for the model (naming model1)

doc1.pdf (54.9 KB)

(Note - here you can assume x1, x2 and x3 to be significant)

b) Then I do the predictions using the above model (naming model1) for the test data.

Doc2.pdf (34.1 KB)

Result - Here I am getting the predicted values of all the data points.

2. For Factor Variable (both normal and ordered Factor variable)-

a) I factor the x1 variable and the run the logistic regression on the training data.
I get the below summary for the model (naming model2).

Doc3.pdf (57.5 KB)

Note - Now assume that in x1 factored variable one of the level is insignificant eg x12 (variable x1 with level 2) and rest variables are significant.

b) Then I run the model again (naming model3), summary is below -

Doc5.pdf (56.3 KB)

c) Now I do the prediction using the test data using the above model (model3), then I get the following summary.

doc 6.pdf (32.3 KB)

Result - Here I am getting the not getting the predicted values for some of the data points, it showing NA.

Main Question - Can anyone tell me how should I exclude or deal with the insignificant levels of a factored variable or ordered factor variable in regression model or any other model and after excluding the same how can I predict the variables for all the data points ?

Thanks and Regards,
Karan Sehgal

[pieterjanvc]

Hi,

Although you made a great effort to try and explain the issue, could you create a reprex for us where we can actually run the code with a minimal dataset you create. I still don't understand what you exactly mean by removing an insignificant factor (do you mean removing a whole feature i.e. column, or just one of the options within a factor..)

This link will help create a reprex:

FAQ: How to do a minimal reproducible example ( reprex ) for beginners

A minimal reproducible example consists of the following items: A minimal dataset, necessary to reproduce the issue The minimal runnable code necessary to reproduce the issue, which can be run on the given dataset, and including the necessary information on the used packages. Let's quickly go over each one of these with examples: Minimal Dataset (Sample Data) You need to provide a data frame that is small enough to be (reasonably) pasted on a post, but big enough to reproduce your issue. Let's say, as an example, that you are working with the iris data frame head(iris) #> Sepal.Length Sepal.Width Petal.Length Petal.Width Species #> 1 5.1 3.5 1.4 0.…

Kind regards,
PJ