Multiple CRF-Extractors

Hello Community,

I would like to know if it is possible to use two CRFEntityExtractors, with different configurations.

The problem ist, I have different entities, which needs different textfeatures for the recognition.

On the first Entity there is an overfitting between test and training data. On the other entity the model generalizes correctly.

I would like to use different textfeatures for the recognition of one entity as I would use for the other entity. Because I think the overfitting is caused by using irrelevant features for training. Is it possible to use the same module with different configurations?



You can run another CRF-entity extractor on a NLU server and talk to it via REST API within a custom component.