After doing some research I couldn't really find a clear difference between those two terms when considering machine learning except that a covariate has to be continuous.
$\begingroup$
$\endgroup$
1
-
2$\begingroup$ Covariates/features are synonymous. The former is preferred by statisticians, while the latter is preferred by the machine learning community $\endgroup$– Demetri PananosCommented Dec 23, 2020 at 17:07
Add a comment
|
1 Answer
$\begingroup$
$\endgroup$
There's a lot of confusion of terms in moving between statistics and machine learning. These two terms are essentially synonyms.
Covariate implies a separate, related, focal variable (e.g., the dependent variable in regression). You can talk about features without a focal variable in mind. But sometimes people talk about covariates without a specified dependent variable, and there's nothing wrong with calling your independent regression variables "features".
a covariate has to be continuous
...that is not true.
