Is "accuracy" the same as "AUC" for a logistic model?

M1. The Predictive Modeling Pipeline M1. Working with numerical data
#1

Hi,

In Module M1
Working with numerical data
I read:
“Accuracy of logistic regression: 0.807”

I was wondering what is the relationship between what is called “accuracy” and what is called an “AUC” (Area Under the ROC curve) for a logistic model?
https://www.graphpad.com/guides/prism/latest/curve-fitting/reg_logistic_roc_curves.htm

I know how an AUC is obtained for a “classical logistic regression model” in SPSS or R (not via Machine Learning) but how is “accuracy” calculated in, for instance, Module 1?

Thank you in advance for your answer.

Best wishes,
Claudine

#2

Hi Claudine,

Accuracy is possibly the most widely used summary metric as its main advantage is a natural interpretation: the proportion of correctly classified samples. However, it is misleading when the data is imbalanced and does not provide any information if the classifier is making errors in the form of false positives or false negatives.

We will discuss these topics in more detail in Module 7.

Best,
Arturo.