Need help on using Metacost operator
ram_nit05
New Altair Community Member
Hi,
Am trying to penalize the misclassification of rare categories in my model by using Metacost operator. However am facing some difficulty in including it in my current flow which starts with examplesource, and then has %xvalidation with libsvmlearner and operatorchain(with both model applier and classification performance).
If I include it within %xvalidation after libsvmlearner it says that the operator cannot take more than two nodes, and if I have Metacost in the operator chain before model applier, I get an error that it cannot take numerical attributes.
Can anybody help me with some link or example where I can get information on how to use this operator?
Thanks,
Ram
Am trying to penalize the misclassification of rare categories in my model by using Metacost operator. However am facing some difficulty in including it in my current flow which starts with examplesource, and then has %xvalidation with libsvmlearner and operatorchain(with both model applier and classification performance).
If I include it within %xvalidation after libsvmlearner it says that the operator cannot take more than two nodes, and if I have Metacost in the operator chain before model applier, I get an error that it cannot take numerical attributes.
Can anybody help me with some link or example where I can get information on how to use this operator?
Thanks,
Ram
Tagged:
0
Answers
-
Try taking a look at 02_Learner\18_SimpleCostSensitiveLearning.xml in the RM Samples directory. It shows an example of how to combine XValidation with MetaCost.
One tip: if you need to do two steps in a place that only allows one node, put them both inside an OperatorChain node. To the node it's embedded in, it will look like a single node, even if you have 5 different steps inside of it.
Keith
0 -
Thanks Keith,
Ramkumar0