Creating numerical variables from categorical variables in an unbalanced dataset

Hello there,

I would like to apply Random Forrest method in a highly unbalanced dataset that includes both numerical and categoorical variables.In order to improve my classification results, before applying the method for classification I thought to create synthtic datasets using the SMOTE and the ADASYN algorithm. However, both methods work only with numerical variables, therefore, I would like to ask if you have any suggestion regarding the way to transform my categorical variables into numerical ones.

With many thanks in advance for your help

Answer 1

[Lei Hou]

Hi Grigorios,

You can do something as the following.

catVar = categorical(["a" "b" "c" "b" "a"]);
numValue = [0.1 3 100]; % The order of numbers refers to the order of categories returned by categories(catVar)
numVar = numValue(catVar)

Hoping my solution helpful to you.

[Grigorios Kyritsakas]

Thanks for your help