Automatic Feature Type Discovery

[jichao06]

Is there plan to support text features in addition to numeric and categorical? For example, apply CountVectorizer then TfIdfTtansformer? If that is intentionally left for user, what is the reason?

Is there plan to support automatic feature type discovery? Basically, instead of relying on user to specify the type of features, infer that info directly from data.

[mfeurer]

The aim of our project is to automate the classifier/regressor prediction pipeline of scikit-learn. If CountVectorizer and TFIDFTransformer work within a pipeline we can put it on our feature request list. Please provide me with an example then.

If scikit-learn implements automatic feature type discovery we can provide that as well. Do you have any reference for that? Also, it would probably be best to request this feature on the scikit-learn mailing list.

[github-actions]

This issue has been automatically marked as stale because it has not had recent activity. It will be closed if no further activity occurs for the next 7 days. Thank you for your contributions.

[eddiebergman]

The text processing part of this issue is being addressed with PR #1300

[mfeurer]

For reference, sklearn thinks about adding a transformer for date types: scikit-learn/scikit-learn#21968