Apologies if this is a very obvious question, but I have been reading various posts and can't seem to find a good confirmation. In the case of classification, is a classifier's accuracy = 1- test error rate? I get that accuracy is $\frac{TP+TN}{P+N}$, but my question is how exactly are accuracy and test error rate related.
$\begingroup$
$\endgroup$
asked Jan 14, 2015 at 21:47
Add a comment
|
2 Answers
$\begingroup$
$\endgroup$
1
In principle, accuracy is the fraction of properly predicted cases.
This is the same as 1 - the fraction of misclassified cases or 1 - the *error* (rate).
Both terms may be sometimes used in a more vague way, however, and cover different things like class-balanced error/accuracy or even F-score or AUROC -- it is always best to look for/include a proper clarification in the paper or report.
Also, note that test error rate implies error on a test set, so it is likely 1-test set accuracy, and there may be other accuracies flying around.
-
$\begingroup$ Yeah, I think that is the issue I was having is that the terms are used vaguely, and you make a good point that it must be reported in the context of your analysis. Thanks for clarifying! $\endgroup$ Commented Jan 16, 2015 at 17:57
$\begingroup$
$\endgroup$
@mbq answered:
"1-the fraction of misclassified cases, that is error(rate)"
However, it seems wrong as misclassification and error are the same thing. See below (from http://www.dataschool.io/simple-guide-to-confusion-matrix-terminology/):
Accuracy: Overall, how often is the classifier correct? (TP+TN)/total = (100+50)/165 = 0.91
Misclassification Rate: Overall, how often is it wrong? (FP+FN)/total = (10+5)/165 = 0.09 equivalent to 1 minus Accuracy
also known as "Error Rate"
