Data Smoothing in Data Cleaner App.

Hello everyone, I have a dataset with 28 features and 508 instances, and I’d like to apply data smoothing to see whether it makes any difference in the results generated by the Classifier Learner app. Some of my variables, such as age, are stored as double even though they have integer values. Should I exclude these variables when using the Data Cleaner app because smoothing age values didn't make much sense? Thanks for the help.

[dpb]

Agreed, one certainly wouldn't want to smooth ages across observations -- or any other variables that are inherent to the observation. It wouldn't be clear what other corallary variables might have been collected for which smoothing would be reasonable and it would only seem at all reasonable if there were a way to do something along a time or other sampled variable and even then by individual--across all individuals wouldn't seem appropriate, either.

But, without knowing much more about what these observations are and how were collected, it's not possible to provide real advice, only some reaction.

[MByk]

Thanks for the answer. I will exclude age, categorical variables, and other features that don’t make sense to smooth.

[dpb]

Again, without knowing anything more than the little we've been told, I'd expect any smoothing should still only be by the individual and not as a whole unless it is some global environmental variable to which all are subject simultaneously.