The logic of Vector Linear Regression

Hallo,

Please, have a look at these two videos produced by Ingo, one of the most senior persons from Rapidminer:

https://www.youtube.com/watch?v=maCLzxLyX6U

https://www.youtube.com/watch?v=YsiWisFFruY

With a quick search along "Vector Linear Regression" & "Linear Regression", it is easy to find more formal (mathematical) explanation (algebra, vector theory).

Bonne chance,

Maerkli

"Ingo, one of the most senior persons from Rapidminer"

FYI @IngoRM is one of two founders of RapidMiner (the other is @RalfKlinkenberg) and they literally wrote the first versions of RapidMiner back 10 years ago (it was originally called YALE). So yeah, @IngoRM is definitely one of our more senior members!

Scott

@sgenzer the operative word is "senior!" hehehe!

Thank you for the response.

But as you know support vector machines (and SVR) are something different than vector linear regression.

So, any difference between vector linear regression and support vector regression?

Hi,

as far as i know: Vector linear regression is a simple linear regression but with many labels. these are specified by having label_1 etc as role.

Best.

Martin

Thank you,

I implemented both linear regression and vector linear regression on a single labeled dataset, but the results were totally different!

So, any other idea about the differences of these two algorithms?

Hi,

did you deactivate everything fance e.g. removal of collinear features in the LinReg?

Best,

Martin

Thank you, it worked!

I unchecked the 'eliminate colinear features' and changed the feature selection to 'Iterative T-Test'. Now the linear regression and vector linear regression are giving same results.