View Single Post
  #2  
Old May 11th, 2008, 05:45 PM
galactus's Avatar
galactus galactus is offline
Eater of Worlds

 
Join Date: Jul 2006
Location: Chaneysville, PA
Posts: 2,990
Country:
Thanks: 130
Thanked 1,146 Times in 1,031 Posts
galactus has much to be proud ofgalactus has much to be proud ofgalactus has much to be proud ofgalactus has much to be proud ofgalactus has much to be proud ofgalactus has much to be proud ofgalactus has much to be proud ofgalactus has much to be proud ofgalactus has much to be proud of
Default

It's called a linear regression. And your data results in a 'not too accurate' one.

A quartic regression would be better.

Using your data I get a quartic of

y=.005286x^{4}-.086979x^{3}+.303854x^{2}+.807917x+25.47

This has an R^2 of 1 and that is as good as it gets.

When doing a regression, the closer R^2 is to 1, the better.

A linear regression resulted in about .4. That is bad.

You can do these with Excel or any decent calculator.
Reply With Quote