https://communities.sas.com/t5/Statistical-Procedures/GENMOD-log-link-equation/m-p/700220#M33767 <P>When GENMOD applies a log link to gamma-distributed y, it is making y into natural* log of y? What is the equation behind the log transfomation?</P> <P>*just found out that Log function in sas means ln (vs. Log10, which is log base 10)</P> <P>&nbsp;</P> <P>&nbsp;</P> Thu, 19 Nov 2020 15:43:23 GMT pink_poodle 2020-11-19T15:43:23Z https://communities.sas.com/t5/Statistical-Procedures/GENMOD-log-link-equation/m-p/700220#M33767 <P>When GENMOD applies a log link to gamma-distributed y, it is making y into natural* log of y? What is the equation behind the log transfomation?</P> <P>*just found out that Log function in sas means ln (vs. Log10, which is log base 10)</P> <P>&nbsp;</P> <P>&nbsp;</P> Thu, 19 Nov 2020 15:43:23 GMT https://communities.sas.com/t5/Statistical-Procedures/GENMOD-log-link-equation/m-p/700220#M33767 pink_poodle 2020-11-19T15:43:23Z https://communities.sas.com/t5/Statistical-Procedures/GENMOD-log-link-equation/m-p/700246#M33770 <P>No, it doesn't make Y into log(Y). It says that Y = exp(b0 + b1X) + error.</P> <P>These are similar, but differ in the way that the errors are modeled. For a complete comparison, discussion, and graphs, see <A href="https://blogs.sas.com/content/iml/2015/09/16/plot-distrib-exp.html" target="_self">"Error distributions and exponential regression models."</A></P> Thu, 19 Nov 2020 16:09:47 GMT https://communities.sas.com/t5/Statistical-Procedures/GENMOD-log-link-equation/m-p/700246#M33770 Rick_SAS 2020-11-19T16:09:47Z