Evaluation Measures

Jump to bottom

janvanrijn edited this page Aug 15, 2018 · 2 revisions

The following evaluation measures are currently calculated by the Evaluation Engine: Regression:

mean_absolute_error
mean_prior_absolute_error
number_of_instances
root_mean_squared_error
root_mean_prior_squared_error
relative_absolute_error
root_relative_squared_error

Classification:

average_cost (based on cost matrix)
total_cost (based on cost matrix)
mean_absolute_error
mean_prior_absolute_error
root_mean_squared_error
root_mean_prior_squared_error
relative_absolute_error
root_relative_squared_error
prior_entropy
kb_relative_information_score
predictive_accuracy
kappa
number_of_instances
precision (per class)
recall (per class)
f_measure (per class)
area_under_roc_curve (per class)
confusion_matrix