0

I have a data frame in csv format. When I try to perform OLR(Ordinal logistic regression) on the data in R, I get this error. I am trying to do credit risk rating so I am including all of the variables in the regression although I know there is multicollinearity between some of them.(I want to see which ones predict better and then eliminate the rest) Here is my code:

modelOLR2 <- polr(as.factor(Aging) ~ Amount + FICO + DTI + DTHC + 
  Years_Between + Delinquencies + Collection_Number + 
  Collection_Amount + Inquiries + Mortgage + Installments + 
  Additional + Bankruptcy + Total_Debt,  
  data=training_set, Hess=TRUE)

modelOLR2 <- polr(as.factor(Aging) ~ .- Credit_Risk_Rating-
  CRR_Points-Housing-Employment-FICO-Housing,  
  data=training_set, Hess=TRUE)

summary(modelOLR2)

I tried both models and they both give the error "Error in svd(X) : infinite or missing values in 'x'"

I already replaced NAs with the average scores for each variable.

Does anyone have any suggestions for me?

DaveArmstrong
  • 11,680
  • 1
  • 10
  • 21
user16452163
  • 1
  • Welcome to SO! You will be most likely to get your question answered if you follow the [helpful advice](https://stackoverflow.com/help/how-to-ask) about asking questions and create a [reproducible example](https://stackoverflow.com/questions/5963269/how-to-make-a-great-r-reproducible-example). – DaveArmstrong Sep 23 '21 at 23:23

0 Answers0