Problem running XGboost with unballanced data

Hi,

I’m using SMOTE node to balance a set of examples. When I use the XGBoost Predictor to use the model trained on the balanced data I get the following error:
WARN XGBoost Predictor 0:20:929 The column ‘EdadCompensar (Non-Native [interface org.knime.core.data.DoubleValue, interface org.knime.core.data.LongValue, interface org.knime.core.data.DataValue, interface org.knime.core.data.ComplexNumberValue, interface org.knime.core.data.FuzzyNumberValue, interface org.knime.core.data.FuzzyIntervalValue, interface org.knime.core.data.BoundedValue])’ is not compatible with the required column ‘EdadCompensar (Number (double))’.
with the following workflow:

tried to use a column rename before the predictor to force the columns to double but it didn’t have any effect.

It is possible that the rename node is not the best way to convert values. There was this suggestion from the forum about SMOTE and Integer variables, you might give that a try:

2 Likes

Thanks for your suggestion. It works now

1 Like

This topic was automatically closed 7 days after the last reply. New replies are no longer allowed.