JohnDeJesus22

recall = cm[1,1]/(cm[1,1] + cm[1,0])
specificity = cm[0,0]/(cm[0,0] + cm[0,1])
balanced_accuracy = (recall + specificity)/2

JohnDeJesus22

numerator = cm[1, 1]
denominator = cm[1, 1] + cm[1, 0] + cm[0, 1]
critical_success_index = numerator/denominator

JohnDeJesus22

numerator = cm[1, 0]
denominator = cm[1, 0] + cm[0, 0]
false_omission_rate = numerator/denominator

JohnDeJesus22

numerator = cm[0, 1]
denominator = cm[0, 1] + cm[1, 1]
false_discovery_rate = numerator/denominator

JohnDeJesus22

numerator = cm[1, 0]
denominator = cm[1, 0] + cm[1, 1]
false_negative_rate = numerator/denominator

JohnDeJesus22

numerator = cm[1,0]
denominator = cm[0,1] + cm[0, 0]
false_positive_rate = numerator/denominator

JohnDeJesus22

numerator = cm[0,0]
denominator = cm[0,0] + cm[1, 0]
negative_predictive_value = numerator/denominator

JohnDeJesus22

numerator = cm[1, 1]
denominator = cm[1, 1] + cm[0, 1]
precision = numerator/denominator

JohnDeJesus22

numerator = cm[0, 0]
denominator = cm[0, 0] + cm[0, 1]
specificity = numerator/denominator

JohnDeJesus22

numerator = cm[1, 1]
denominator = cm[1,1] + cm[1, 0]
recall = numerator/denominator