CODEWITHSUNDEEP
pythonfunctionparametersscikit-learnevaluation
Social Programmer
Social Programmer

Reputation: 167

sklearn , change default parameters in cross_val_score

I am evaluating text classification predictions , with cross_val_score. I need to evaluate my predictions with recall_score function , but with parameter average = 'macro'. cross_val_score sets it to the default parameter , binary , which doesnt work to my code. Is there any way to call recall_score with a different parameter , or change the default parameter to macro.

results = model_selection.cross_val_score(estimator, X, Y, cv= kfold, scoring= 'recall')

Upvotes: 1

Views: 1520

Answers (1)

Vivek Kumar
Vivek Kumar

Reputation: 36599

You can just use "recall_macro" in it like this:

results = model_selection.cross_val_score(estimator, X, Y, cv= kfold, scoring= 'recall_macro')

According to the documentation of metrics

‘f1’                metrics.f1_score            for binary targets
‘f1_micro’          metrics.f1_score            micro-averaged
‘f1_macro’          metrics.f1_score            macro-averaged
‘f1_weighted’       metrics.f1_score            weighted average
‘f1_samples’        metrics.f1_score            by multilabel sample
‘neg_log_loss’      metrics.log_loss            requires predict_proba support
‘precision’ etc.    metrics.precision_score     suffixes apply as with ‘f1’
‘recall’ etc.       metrics.recall_score        suffixes apply as with ‘f1’

As you can see, its specified that all suffixes apply to "recall".

Alternatively, you can also use make_scorer like this:

# average can take values from 'macro', 'micro', 'weighted' etc as specified above
scorer = make_scorer(recall_score, pos_label=None, average='macro')
results = model_selection.cross_val_score(estimator, X, Y, cv= kfold,
                                          scoring= scorer)

Upvotes: 3

Related Questions