Turn on suggestions

Auto-suggest helps you quickly narrow down your search results by suggesting possible matches as you type.

Showing results for

- JMP User Community
- :
- Discussions
- :
- Discussions
- :
- Quadratic Regression Fit - Please Help

Topic Options

- Start Article
- Subscribe to RSS Feed
- Mark Topic as New
- Mark Topic as Read
- Float this Topic for Current User
- Bookmark
- Subscribe
- Printer Friendly Page

- Mark as New
- Bookmark
- Subscribe
- Subscribe to RSS Feed
- Get Direct Link
- Email to a Friend
- Report Inappropriate Content

Nov 12, 2015 10:46 AM
(5120 views)

Hello,

When I perform a quadratic regression, it returns the form y=a+bx+c(x-d)^2. I would like it in the form y=a+bx+cx^2 so that I can have the correct parameter estimates. How can I do this - either change the way the model returns or change the parameters as needed?

Thanks,

Brian

1 ACCEPTED SOLUTION

Accepted Solutions

- Mark as New
- Bookmark
- Subscribe
- Subscribe to RSS Feed
- Get Direct Link
- Email to a Friend
- Report Inappropriate Content

Hi, Brian!

Try using the Fit Special... dialog of the Fit Y by X platform (Under the LIRT by Bivariate Fit).

Leave both Y and X as No Transformation, specify the degree as 2 Quadratic, and uncheck the Centered Polynomial checkbox. That will give you the model you seek.

As an aside, JMP almost always centers the variables in analyses since it makes a lot of statistical sense (makes parameters comparable, etc.).

As another, more distant, aside LIRT stands for Little Inverted Red Triangle.

2 REPLIES 2

- Mark as New
- Bookmark
- Subscribe
- Subscribe to RSS Feed
- Get Direct Link
- Email to a Friend
- Report Inappropriate Content

Hi, Brian!

Try using the Fit Special... dialog of the Fit Y by X platform (Under the LIRT by Bivariate Fit).

Leave both Y and X as No Transformation, specify the degree as 2 Quadratic, and uncheck the Centered Polynomial checkbox. That will give you the model you seek.

As an aside, JMP almost always centers the variables in analyses since it makes a lot of statistical sense (makes parameters comparable, etc.).

As another, more distant, aside LIRT stands for Little Inverted Red Triangle.

- Mark as New
- Bookmark
- Subscribe
- Subscribe to RSS Feed
- Get Direct Link
- Email to a Friend
- Report Inappropriate Content

Re: Quadratic Regression Fit - Please Help

Thanks Kevin! Exactly what I was looking for!