Interpreting outputs of calc.relimp

General
1

I am using the package "relaimpo" to calculate the relative importance of variables that I have fitted to a linear model. Code is:

calc.relimp(model,rela=TRUE)

It is returning me outputs like these:

> calc.relimp(model,rela=TRUE)
Response variable: Sales 
Total response variance: 3505.987 
Analysis based on 20 observations 

2 Regressors: 
Price Advertising 
Proportion of variance explained by model: 53.51%
Metrics are normalized to sum to 100% (rela=TRUE). 

Relative importance metrics: 

                  lmg
Price       0.2246168
Advertising 0.7753832

Average coefficients for different model sizes: 

                   1X       2Xs
Price       -16.92037 -21.71100
Advertising  68.88305  74.38533

I understand that ultimately I just need to look at the relative importance metrics for both variables (ie 22.5% for Price and 77.5% for Advertising). But what do the numbers in "Average coefficients for different model sizes" mean? Thanks.

2

Can you provide a reproducible example? FAQ: What's a reproducible example (`reprex`) and how do I do one?

3

I ran reprex and got the following:

calc.relimp(model,rela=TRUE)
#> Error in calc.relimp(model, rela = TRUE): could not find function "calc.relimp"

Created on 2020-08-05 by the reprex package (v0.3.0)

Not sure if this is helpful.

Basically, I'd just like to get an understanding of the numbers under "1X" and "2Xs" in the calc.relimp output.

4

Not really, but this is a description:

average coefficients for variables not not requested by always only for models of different sizes;

note that coefficients refer to modeling residuals after adjusting out variables listed in always (both from response and other explanatory variables)

From here:
https://rdrr.io/cran/relaimpo/man/calc.relimp.html

And a really detailed description of the function here:

5

This topic was automatically closed 21 days after the last reply. New replies are no longer allowed.