I know that for an lm model is enough to run plot(model_lm) to get diagnostic plots. I am dealing with high-dimensional count data, so I am using Poisson, Quasipoisson, Negative Binomial and Zero Inflated . Is there a way to compare different glm outputs to see which one fit the data better? Moreover, how can I plot diagnostic plots?

My question is a bit different from: Diagnostic plots for poisson regression(GLM) since my X is made of five numeric variables.

  • 1
    $\begingroup$ Answers to the duplicate explicitly include multiple regressions (more than one explanatory variable). $\endgroup$ – whuber Mar 13 '19 at 19:12