I'm wondering what is the estimation techniques used in lm(). If it's OLS, how could we perform a log likelihood test by logLik()? What's the difference between lm() and ols(), mle() and other functions?
Asked
Active
Viewed 4,187 times
3
lmuses OLS. The basic algorithm (the same one used in most packages) is described here. There's noolsormlefunction that I could find in the distribution of R; there may be more than one such function elsewhere -- what packages are you talking about? – Glen_b Aug 16 '14 at 09:37olsis part of thermspackage. It apparently provides better support for categorical variables, but mostly breaks compatibility with any functions expectinglmoutput. Not surprising in this duct-tape universe of R. – jessexknight Feb 04 '20 at 16:44