Evaluation

model.plot('metric_name')

There are several different evaluation metrics you can choose from for regression and classification tasks. They are derived from Scikit-Learn’s own metrics. All you need to is type in the metric name in string format when using the .plot method.

Regression metrics

Parameter	String format
explained variance	‘explained_variance’
max error	‘max_error’
mean absolute error	‘mean_absolute_error’
mean squared error	‘mean_squared_error’
mean squared log error	‘mean_squared_log_error’
median absolute error	‘median_absolute_error’
r2 score	‘r2_score’
mean absolute error score	‘mean_absolute_error_score’
d2 absolute error	‘d2_absolute_error_score’
d2 pinball score	‘d2_pinball_score’
d2 tweedie score	‘d2_tweedie_score’

Classification metrics

Parameter	String format
accuracy score	‘accuracy_score’
balanced accuracy score	‘balanced_accuracy_score’
f1 score	‘f1_score’
precision score	‘precision_score’
recall score	‘recall_score’
jaccard score	‘jaccard_score’
roc auc score	‘roc_auc_score’
log loss	‘log_loss’
average precision score	‘average_precision_score’
brier score loss	‘brier_score_loss’