CODEWITHSUNDEEP
roversampling
Cigdem
Cigdem

Reputation: 1

Error in Rose sampling when balancing data with categorical variables

I'm trying to balance my data in which the majority class has proportion of 99% while the rare class has 1%. My response variable is binary and my independent variables are both binary, integer and categorical variables.

I'm using ROSE function of ROSE package but I am getting the error below: "Error in rose.sampl(n, N, p, ind.majo, majoY, ind.mino, minoY, y, classy, : The current implementation of ROSE handles only continuous and categorical variables."

I understand the error is because of my independent variables but I could not figure out what the main reason for this error is and how I can solve it. Do you have any experience with such an error?

Upvotes: 0

Views: 2891

Answers (1)

katem
katem

Reputation: 51

I had a similar message. Turns out one of my independent variables was a difftime object when it should've been made numerical.

The ROSE package's documentation states it accepts continuous variables in integer or numerical format, and categorical variables in factor format, for its accuracy.meas and roc.curve functions. It's likely the same for the ROSE function. Knowing the classes of your variables would've been helpful, but consider factorising those binary columns of yours or making sure your other variables adhere to these data types.

Upvotes: 4

Related Questions