The Altair Community is migrating to a new platform to provide a better experience for you. In preparation for the migration, the Altair Community is on read-only mode from October 28 - November 6, 2024. Technical support via cases will continue to work as is. For any urgent requests from Students/Faculty members, please submit the form linked here
"Applying
Maven
Hi @ all!
I have some data with a lot of numerical values. I want to performe some kind of feature selection. Therefore I discretize (Discretize by Entropy) the data and then I run some attribute selection operator (selfmade). This works out just fine. But for learning from my data I want to use the numerical values not the discretized ones.
So my operator provides an output port which deliveres the results of the feature selection as attribute weights (values only 1 and 0). The question is now how to apply those weights to my initial data? The simplest way (loading the original data, using Scale by Weight operator) doesn't work, as attributes which are not in the weight vector aren't touched. So the final example set contains all selected features and all features which were eliminated by the Entropy Discretizing operator (it automatically eliminates all attributes having only one discretized value -- which is highly appreciated).
Does anybody have an idea?
Best regards,
Michael
I have some data with a lot of numerical values. I want to performe some kind of feature selection. Therefore I discretize (Discretize by Entropy) the data and then I run some attribute selection operator (selfmade). This works out just fine. But for learning from my data I want to use the numerical values not the discretized ones.
So my operator provides an output port which deliveres the results of the feature selection as attribute weights (values only 1 and 0). The question is now how to apply those weights to my initial data? The simplest way (loading the original data, using Scale by Weight operator) doesn't work, as attributes which are not in the weight vector aren't touched. So the final example set contains all selected features and all features which were eliminated by the Entropy Discretizing operator (it automatically eliminates all attributes having only one discretized value -- which is highly appreciated).
Does anybody have an idea?
Best regards,
Michael
Tagged:
0
Answers
Just after I posted my message I found the solution.
For all who have a similar problem, the solution is: Use Select by Weight operator. It has a parameter to deselect unweighted attributes.
Best regards,
Michael