[PYTHON] Setting of evaluation index for cross-validation (memo, scikit-learn, cross_validation.cross_val_score)

case of cross_validation.cross_val_score

Specify the value of this ** scoring **! !!

scores = cross_validation.cross_val_score(clf, X_digits, y_digits, cv=k_fold, ... scoring='precision_macro')

Even when I heard the presentation at the conference, the evaluation index of cross-validation was often a mystery. I was wondering if it would be uniquely determined depending on the target (old days). You choose properly depending on your purpose ^^;

** List list, details can be jumped further from the link ** http://scikit-learn.org/stable/modules/model_evaluation.html#scoring-parameter

['accuracy', 'adjusted_rand_score', 'average_precision', 'f1', 'f1_macro', 
'f1_micro', 'f1_samples', 'f1_weighted', 'log_loss', 'mean_absolute_error', 
'mean_squared_error', 'median_absolute_error', 'precision',
'precision_macro', 'precision_micro', 'precision_samples', 
'precision_weighted', 'r2', 'recall', 'recall_macro', 'recall_micro', 
'recall_samples', 'recall_weighted', 'roc_auc']

'accuracy'　accuracy_score　 Many samples do so in binary classification, but this is the default? Formally, it's like this.

At the time of the return, Mr. Daibutsu wrote. The table has neg_, but the list says "mean_squared_error". It's related to the fact that negative values come back ... http://d.hatena.ne.jp/teramonagi/20130825/1377434479#20130825f2

** Excerpt from description (quotation is the parameter value) ** The notation exactly matches the list ... It is a mystery that it does not match

Classification ‘accuracy’ metrics.accuracy_score ‘average_precision’ metrics.average_precision_score ‘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’ ‘roc_auc’ metrics.roc_auc_score Clustering ‘adjusted_rand_score’ metrics.adjusted_rand_score Regression ‘neg_mean_absolute_error’ metrics.mean_absolute_error ‘neg_mean_squared_error’ metrics.mean_squared_error ‘neg_median_absolute_error’ metrics.median_absolute_error ‘r2’ metrics.r2_score