In glm() function, it seems that you can specify Normal, Gamma, Inverse Gaussian, Poisson, Binomial, and quasi families. See here: https://stat.ethz.ch/R-manual/R-devel/library/stats/html/family.html. But how do you specify an Exponential distribution?

  • $\begingroup$ The exponential distribution is a Gamma distribution with shape parameter $1$. $\endgroup$ – whuber Dec 6 '16 at 23:11
  • $\begingroup$ Answered here: stats.stackexchange.com/questions/240455/… ... both Ben's answer and mine contain the essential information (his was just after mine but with more details) $\endgroup$ – Glen_b -Reinstate Monica Dec 6 '16 at 23:17
  • $\begingroup$ Thanks! Is there an easy way to get plots and fitted values after summary(glm.fit, dispersion = 1), where glm.fit is the fitted glm? I used to do plot(glm.fit) and fitted.values(glm.fit)... $\endgroup$ – hiya Dec 6 '16 at 23:58