Small Solver Program

[Thomas Klemm]

(02-18-2019 10:22 PM)Thomas Klemm Wrote:  Quadratic fit without linear algebra (32SII) - NOT only for statistics lovers - LONG

(02-19-2019 08:39 AM)Csaba Tizedes Wrote:  But as a programming question this was an interesting "game" for me.

Since the regression is linear you can remove division by 2 at two places in your program.
First when calculating the midpoints:

Code:

RCL X
RCL+U
2
/
Sum+

And then later on when calculating A:

Code:

LBL C     #Calculating A, B and C
m
2
/
STO A

We can rewrite \(y=m\frac{x_k+x_{k+1}}{2}+b\) as \(y=\frac{m}{2}(x_k+x_{k+1})+b\).
But since \(A=\frac{m}{2}\) this becomes \(y=A(x_k+x_{k+1})+b\).

Cheers
Thomas