Started By
Message

re: Excel help

Posted on 9/15/15 at 1:40 pm to
Posted by K E V 8 4
Member since Jul 2010
608 posts
Posted on 9/15/15 at 1:40 pm to
I think I might see what's happening. A four-order poly gives the following (and the line is perfectly through the points):

y = 15297x4 - 457991x3 + 5E+06x2 - 3E+07x + 5E+07
R² = 0.9999

The x4, x3, etc. are actually x^4, x^3, etc.

Yet.....I plugged in 6.27 and get some nonsensical answer- negative 30.8 million. I think the XLS reported coefficients are not providing enough significant digits and that is what's messed up. Note the 5.00 million, 30.0 million and 50.0 million coefficients. The rounded (truncated) diffs is what's busting the equation. Would need to do the calcs manually (or program) to get the precise coefficients. An XLS bug, it appears. Somebody shout if that's not it.
Posted by K E V 8 4
Member since Jul 2010
608 posts
Posted on 9/15/15 at 1:43 pm to
Solution: divide all of your Y data points by 10,000. Regress. Then take result of equation and multiply by 10,000. The equation and RSq are as follows:

y = 1.5297x4 - 45.799x3 + 515.07x2 - 2580.4x + 4862.3
R² = 0.9999



Answer: 71,386
This post was edited on 9/15/15 at 1:51 pm
first pageprev pagePage 1 of 1Next pagelast page
refresh

Back to top
logoFollow TigerDroppings for LSU Football News
Follow us on Twitter, Facebook and Instagram to get the latest updates on LSU Football and Recruiting.

FacebookTwitterInstagram