On Tuesday, Mar 30, 2004, at 21:56 US/Pacific, John Pearson wrote:

> I am not very good at math. Please assume I am younger than 9 or older
> than 40.
>
> I want my mpu to be able to create values for a linear regression
> equation from 4 values that are entered, example:
>
> When A = 6, B = 2.1
> When A = 30, B = 0
>
Do you mean "linear regression" (which, being over 40, I don't remember
what that means), or are you just looking for a 'curve fit' so that you
can derive an algorithm or equation so that you can find B if you know
A is 15?  (for example.)

In the latter case, you have a nice straight line equation (since you
have only two points.)  these have the form B = m*A + I where "m" is
the slope of the line ((B1-B2)/(A1-A2)) and I is the "intercept" value
when A is 0.  You can get that pretty easily once you have m and a
known point.  (hmm.  For a computer, re-arrange:)

A1 = m * B1 + I
I = A1 - (m*B1)

so in your case, you have

m = (30-6)/(0-2.1) = -11.43 (more or less)

and then 30 = -11.43*0 + I,  or I = 30 - (-11.43 * 0),  so I = 30.

So you have:  A = -11.43*B + 30

Checking your other point, 6 = -11.43*2.1 + 30?  Yep!

See also:

http://www.mcgrawhill.ca/school/booksites/
calculus+and+advanced+functions/student+resources/toc/
review+of+prerequisite+skills/equation+of+a+line+given+two+points/
worked+example.php

There are more complicated mechanisms for fitting a more complicated
curve to more than two data points, given only the numeric values of
the points, but at that point you should probably seek out a book on
"Numeric analysis."

BillW

--
http://www.piclist.com hint: To leave the PICList
mailto:piclist-unsubscribe-request@mitvma.mit.edu