0
$\begingroup$

Apologize in advance for there being not a lot of wisdom here..

But If I am creating a multiple linear regression model with time series data, (more specifically using day-of-week dummy variables for my independent variables) I would need to build one of my day-of-week variables into the intercept of the model, right?

For example,

if the date = Monday, the value for this variable/column in the dataset would be 1 else 0. if the date = Tuesday, the value for this variable/column in the dataset would be 1 else 0.. if the date = Wednesday, the value for this variable/column in the dataset would be 1 else 0...

And so on for all days of the week. If this is true could someone help explain the theory behind why one independent variable needs to be built into the intercept? I cant remember where I learned this from..

Improve this question
$\endgroup$
3
  • $\begingroup$ You don’t have to build it into the intercept, but if you don’t, leave out an intercept term or your model matrix will not have full rank. You may be interested in a post of mine: stats.stackexchange.com/a/439818/247274 $\endgroup$
    – Dave
    Commented Mar 18, 2020 at 17:48
  • $\begingroup$ Can you explain that more in more detail? It sort of makes sense... $\endgroup$
    – bbartling
    Commented Mar 18, 2020 at 17:50
  • $\begingroup$ I’ve included a link to a post of mine that goes through this. $\endgroup$
    – Dave
    Commented Mar 18, 2020 at 17:52

0

Reset to default

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Browse other questions tagged or ask your own question.