Why does mlbox recognize my multi-class training set as a regression problem
gaoexingyun opened this issue · comments
I originally wanted to train my multi-class training set with mlbox classification, but after importing the data, it automatically recognized it as regression, and changed my target value data type from int to float.
Hello @gaoexingyun,
Thanks for the issue ! You can refer to this one : #87
To sum it up, MLBox always tries to cast your target into float so if you have something like : "1", "2", ... it is considered as a regression problem, better if your target looks like "C1", "C2", ...
Please tell me if it works now.
Hello @gaoexingyun,
Thanks for the issue ! You can refer to this one : #87To sum it up, MLBox always tries to cast your target into float so if you have something like : "1", "2", ... it is considered as a regression problem, better if your target looks like "C1", "C2", ...
Please tell me if it works now.
Thank you. It works.