Found a work around, I skipped preprocessing step for target/dependent variable and it seems to be working fine. Now the actual vs prediction shows actual values and not normalized ones.
... View more
Hi , I get following error message when trying to select split by field option, i choose less than 5 features but still getting following error. I have a total of 7195 rows in my dataset and 17 features excluding time. I referred the documentation but could not make out why this error is occurring Got this error when 1 feature was selected: Error in 'fit' command: Error while fitting "DensityFunction" model: The number of groups can not exceed 1024; the current number of groups is 7195. Please find detailed information about the number of groups in docs. Got this error when 2 features were selected: Error in 'fit' command: Error while fitting "DensityFunction" model: The number of groups can not exceed 1024; the current number of groups is 5452. Please find detailed information about the number of groups in docs.
... View more
Hi, I am looking for command equivalent to inverse_transform in Splunk. after applying Standard scaler pre-processing , all predicted target values are in standard scaler form. which needs to be converted to original levels as shown below, sc_y = StandardScaler() y_pred = sc_y.inverse_transform(regressor.predict(X_test)) Regards, Shobitha
... View more
In Splunk MLTK, how do I convert Target variables back to original levels after applying Standard Scaler pre-processing method. What is SPL command perform this conversion back to original levels
... View more