V
V
Vlad_Fedorenko2015-01-25 21:32:20
data mining
Vlad_Fedorenko, 2015-01-25 21:32:20

Is there a connection between the coefficients of linear regression and the correlation of individual features with the resulting one?

Suppose we have m dimensions of the form (x1,...xn), all xi -numbers, and m values ​​of the resulting feature yi.
Let's bring all xi to the range [0,1] and calculate the feature coefficients for linear regression.
Is it true that the feature with the highest coefficient in the linear model will have the highest correlation with the resulting feature, and the one with the minimum coefficient will have the lowest correlation?

Answer the question

In order to leave comments, you need to log in

1 answer(s)
A
Andy_U, 2015-01-26
@Andy_U

Covariance yes, correlation no.

Didn't find what you were looking for?

Ask your question

Ask a Question

731 491 924 answers to any question