How to prevent or remove or replace NA values in glm output statistics?

Question

In the following dataset, I had just trained a generalized linear model with gamma-distributed residuals, refer the following code:

library(MASS)

df <- read.csv('test.csv')

model <- glm(formula = y ~ method * site + year + 0,

             family=Gamma(link = "log"), data = df)

And by executing the above code, I am getting something like this:

> summary(model) 

Coefficients:

                  Estimate Std. Error t value Pr(>|t|)

methodM0          3.89533    0.13670  28.496  < 2e-16 ***

methodM1          5.63965    0.20940  26.933  < 2e-16 ***

methodM2        -55.854107  73.982453  -0.755    0.450

methodM3        -55.731730  73.986509  -0.753    0.451

siteS1           -0.002872   0.098226  -0.029    0.977

siteS2            0.060892   0.107795   0.565    0.572

siteS3           -0.016239   0.102258  -0.159    0.874

year              0.030813   0.036743   0.839    0.402

methodM1:siteS1  -0.030616   0.144592  -0.212    0.832

methodM2:siteS1  -0.030632   0.144663  -0.212    0.832

methodM3:siteS1   0.064179   0.145593   0.441    0.659

methodM1:siteS2  -0.146505   0.152012  -0.964    0.335

methodM2:siteS2  -0.039610   0.148024  -0.268    0.789

methodM3:siteS2  -0.202881   0.150406  -1.349    0.178

methodM1:siteS3   NA         NA         NA       NA

methodM2:siteS3   0.081617   0.144040   0.567    0.571

methodM3:siteS3  -0.064155   0.147771  -0.434    0.664

I know that the table is the outcome of made-up numbers, but the main query is that I have an interaction between method M1 and site S3 that give NA. But I want to set up the GLM to remove that interaction after training, or assign those NA values in the model to 0?

Answer 1

I think you use update() function, like the following code:

model1 <- glm(formula = y ~ method * site + year + 0,

             family=Gamma(link = "log"), data = df)

model2 <- update(model1, . ~ . - methodM1:siteS3)

If you are a beginner and want to know more about R then do refer to the R programming tutorial.