machine learning - How do you know if a data set is right for linear regression if it has multiple features? -

if has 1 feature it's easy. graph it.

one of records there looks (18, 15). simple. if have multiple features adds more dimensions graph, right? how can visualize data set , determine whether or not linear regression applicable if can't graph it?

by way, aware there whole cluster of algorithms choose , linear regression might not best fit particular problem. i'm asking i'm learning this perspective not what's best way this perspective.

so linear regression assumes data linear in multiple dimensions. wont possible visualize high dimensional data unless use methods reduce high dimensional data. pca can bringing down 2 dimensions won't helpful.

you should cross validation on model see if getting decent fit on data. if not means linear regression not data. .