# Others transformation

In addition to these mathematical transformation, some variables reflecting the relationship with the target variable can also be derived, such as target positive sample proportion, odds encoding, log-odds encoding, and numerical target mean value

Titanic's target variables Survived as categorical variable, and the"Sex" variable is transformed by target positive sample proportion, odds encoding, and log-odds encoding

 A 1 =file("D://titanic.csv").import@qtc() 2 =A1.groups(Sex;count(Survived==1)/count(~):tar_P) 3 =A1.derive(if(Sex=="female",A2(1).tar_P,A2(2).tar_P):tar_P_Sex,if(Sex=="female",A2(1).tar_P/A2(2).tar_P,A2(2).tar_P/A2(1).tar_P):odds,lg(odds):lg_odds)

A2 Group the samples according to “Sex”, count the proportion of group members rescued

A3 Calculate the target positive sample proportion, Odds encoding, log-odds encoding

The target variable "SalePrice" in the house price data is a numerical variable. The categorical variable "MSZoning" is transformed by the numerical target mean value

 A 1 =T("D://house_prices_train.csv") 2 =A1.groups(MSZoning;avg(SalePrice):tar_mean) 3 =A1.derive(A2(A2.(MSZoning).pos(MSZoning)).tar_mean:MSZoing_tar_mean)

A2 According to the "MSZoing" group, the corresponding target variable mean is calculated

A3 According to the value of each variable, the corresponding target mean is obtained