search

MetOffice / XBTs_classification

Project for the classification of eXpendable Bathy Thermographs
BSD 3-Clause "New" or "Revised" License
4 stars 2 forks source link

Set up scikit-learn pipeline #15

Closed stevehadd closed 4 years ago

stevehadd commented 4 years ago

We have done adhoc classification experiments using a variety of classifiers in SK-learn so far. For producing actual results we should set up a proper pipeline with SK-learn pipeline class: https://scikit-learn.org/stable/modules/generated/sklearn.pipeline.Pipeline.html

with multiple classifiers.

stevehadd commented 4 years ago

As of commit edbd963, we now have scikit-learn pipelliens up and running for DecisiontreeClassifier and MLPClassifier. The other classifiers all have the same interface, so I think this functionality can be considered to be implemented. The next step is to get the pipeline working more efficient, potentially using partial fit or dask (#2).