{smcl} {* *! version 0.1.0 11Aug2021}{...} {viewerjumpto "Syntax" "macroF##syntax"}{...} {viewerjumpto "Description" "macroF##description"}{...} {viewerjumpto "Options" "macroF##options"}{...} {viewerjumpto "Examples" "macroF##examples"}{...} {viewerjumpto "Authors" "macroF##authors"}{...} {...}{* NB: these hide the newlines } {...} {...} {title:Title} {p2colset 5 18 20 2}{...} {p2col :{cmd:macroF} {hline 2}} Compute macroF {p_end} {p2colreset}{...} {marker syntax}{...} {title:Syntax} {p 8 16 2} {cmd:macroF} {ifin} [{cmd:,} {it:options}] {synoptset 20 tabbed}{...} {synopthdr} {synoptline} {synopt :{opth pred:(macroF##pred:str)}} Variable name containing predicted classification coded as 0/1.{p_end} {synopt :{opth true:(macroF##true:str)}} Variable name containing true classification coded as 0/1.{p_end} {synoptline} {marker description}{...} {title:Description} {pstd} {cmd:macroF} Computes the macroF evaluation criterion for multi-class outcomes. {marker options}{...} {title:Options} {phang} {marker pred}{...} {opt pred:(str)} Variable name containing predicted classification coded as 0/1.{p_end} {phang} {marker true}{...} {opt true:(str)} Variable name containing predicted classification coded as 0/1.{p_end} {marker examples}{...} {title:Examples} {pstd}Example: Predict 3 types of insurance. Convert probabilities for each to a single variable with the predicted category. Compute macroF. {phang}{cmd:. use https://www.stata-press.com/data/r17/sysdsn1 } {phang}{cmd:. mlogit insure nonwhite age male } {phang}{cmd:. predict p1 p2 p3 } {phang}{cmd:. gen maxp=max(p1,p2,p3) } {phang}{cmd:. gen pred=. } {phang}{cmd:. replace pred=1 if p1==maxp } {phang}{cmd:. replace pred=2 if p2==maxp } {phang}{cmd:. replace pred=3 if p3==maxp } {phang}{cmd:. tab pred,m } {phang}{cmd:. macroF, pred(pred) true(insure) } {marker authors}{...} {title:Authors} {pmore} Matthias Schonlau {p_end}