Well I noticed that sometimes in the regression problems, log-transforming the target, fitting the model, predicting the dataset then inversing the transformation for the predictions gives some increase in the performance, though not a big one.
You can search more about 'log-transforming the target'.
thanks for sharing your solution. I wander why you decide to use TransformedTargetRegressor?
Well I noticed that sometimes in the regression problems, log-transforming the target, fitting the model, predicting the dataset then inversing the transformation for the predictions gives some increase in the performance, though not a big one.
You can search more about 'log-transforming the target'.