[SOLVED] Normalise the training set; also normalise test set?
I normalise all of the attributes in a training set, including the Labelled attribute [is this recommended or not?]
I train a linear regression model using the training set.
I then apply the model to a dataset (the test set). Does this data set need to have the same normalisations as the training set?
If the Labelled attribute has been normalised in the test set, how do I revert to de-normalised data?
The documentation of the "De-Normalise" operator is empty, and I don't understand it.
Cheers,
Ben
I train a linear regression model using the training set.
I then apply the model to a dataset (the test set). Does this data set need to have the same normalisations as the training set?
If the Labelled attribute has been normalised in the test set, how do I revert to de-normalised data?
The documentation of the "De-Normalise" operator is empty, and I don't understand it.
Cheers,
Ben
0
Answers
Ben.