Tpot: Is there refit to all data when .score is called?

Created on 9 Mar 2018  路  1Comment  路  Source: EpistasisLab/tpot

Is there anything similar to scikit-learn GridSearchCV, refit = True?

Meaning that, when k-fold CV is performed, is all of the data fitted one final time before giving a score on an unseen data.

tpot.score(X_test, y_test)

question
Source
picture ogencoglu

Most helpful comment

For now, tpot don't have this refit function. But there is a workaround for it via the codes below:

best_pipeline = tpot.fitted_pipeline_
best_pipeline.fit(X_test, y_test)
best_pipeline.score(X_test, y_test)
picture weixuanfu on 9 Mar 2018
👍2

>All comments

For now, tpot don't have this refit function. But there is a workaround for it via the codes below:

best_pipeline = tpot.fitted_pipeline_
best_pipeline.fit(X_test, y_test)
best_pipeline.score(X_test, y_test)
weixuanfu picture weixuanfu on 9 Mar 2018
👍2
Was this page helpful?
0 / 5 - 0 ratings

Related issues

jonathanng picture jonathanng  路  3Comments
beijingtl picture beijingtl  路  4Comments
stokjakub picture stokjakub  路  3Comments
rhiever picture rhiever  路  4Comments
omarcr picture omarcr  路  4Comments