Talk:What does clustering tell us

From Personal Science Wiki
Jump to navigation Jump to search

-

How to deal with highly correlated variables?[edit | edit source]

It is called collinearity and there are many tutorials right in the first google search. Though its only a problem for simpler ML algorithms you may want to dig deeper for EDA. - DG

In a way that's what the dimensionality reduction tries to achieve, it takes all the variables and rearranges them in a way by correlation between them, that's why in this case I end up with the three main directions (three groups of variables that are highly correlated with each other, two of those groups are in turn negatively correlated with each other). - Gedankenstuecke (talk) 09:52, 13 June 2022 (UTC)