factor based on deviation from mean
Posted: Wed Mar 14, 2018 10:03 am
I have an equal-weighted index of the z-score for 10 variables. For each variable, I want to associate a constant for that variable's z-score relative to the mean z-score of all other variables. For instance, X1 = 2.5 mean(X2-X10) = 2.0 if abs(X1 - mean(X2-X10)>1 then 0.2, >0.5 then 0.4, etc. Basically I want to reduce the weight of variables that significantly deviate from the mean. Does anyone know a code that will facilitate this?