Well, I am studying Machine Learning and I came up with the terminology 'feature crosses' which for me seems similar to an interaction term. Are they the same?
$\begingroup$
$\endgroup$
3
-
$\begingroup$ Naturally, we wouldn't be able to tell w/o you explaining what "feature crosses" means. $\endgroup$– gunesCommented Aug 16, 2020 at 13:00
-
1$\begingroup$ A feature cross is a synthetic feature formed by multiplying (crossing) two or more features. Crossing combinations of features can provide predictive abilities beyond what those features can provide individually.developers.google.com/machine-learning/crash-course/… $\endgroup$– João Vitor GomesCommented Aug 16, 2020 at 13:16
-
$\begingroup$ @gunes any idea? $\endgroup$– João Vitor GomesCommented Aug 16, 2020 at 15:31
Add a comment
|
1 Answer
$\begingroup$
$\endgroup$
4
A feature cross is a synthetic feature formed by multiplying (crossing) two or more features. Crossing combinations of features can provide predictive abilities beyond what those features can provide individually.
Yes, this is exactly the same with feature interactions.
-
$\begingroup$ do you have any references or may reason how they are the same (interaction term = feature cross? $\endgroup$ Commented Aug 16, 2020 at 16:39
-
$\begingroup$ e.g. pycaret.org/feature-interaction $\endgroup$– gunesCommented Aug 16, 2020 at 17:42
-
$\begingroup$ ... or consult any of the (thousands) of posts here on CV discussing interactions: stats.stackexchange.com/search?q=regression+interaction. $\endgroup$– whuber ♦Commented Aug 16, 2020 at 19:21
-
$\begingroup$ it is surely a better suggestion than mine. $\endgroup$– gunesCommented Aug 16, 2020 at 19:22