HP Forums
Stat2Var bug sketch plot - Printable Version

+- HP Forums (https://www.hpmuseum.org/forum)
+-- Forum: HP Calculators (and very old HP Computers) (/forum-3.html)
+--- Forum: HP Prime (/forum-5.html)
+--- Thread: Stat2Var bug sketch plot (/thread-6120.html)



Stat2Var bug sketch plot - akmon - 04-23-2016 10:11 PM

Hello. Sketch plot is a very interesting function for best fit given some points. I wanted to know how accurate is my finger calculating R^2 stats after sketching. I made a very simple example given these points:
0,0
1,1
2,2
3,3
4,4

After drawing the points I sketch with my finger a sort of line, for example: Y=0,6*X+1
When I go to stats sheet, the first colum represents the linear fit (it´s done automaticaly by HP Y=X), and the second column is the equation I obtained by sketching the line with my finger. It gets unexpected values.

[Image: stats2var_zpst9lxebn5.png]

Moreover the more I press STAT softmenu, the more the values changes:

[Image: stat2var2_zpstfans9jk.png]

Have I discovered a bug?


RE: Stat2Var bug sketch plot - Tim Wessman - 04-25-2016 03:00 PM

I believe so. Note that you still wouldn't get a R^2 value in there because it isn't a true regression but acts as a "user function". It would behave the same way as doing one of the other regresisons, then modifying the equation. The type changes to "user" and has much more limited capability.

A lot of people seem to have the idea that R^2 is simply a collection of the error and can be done for any regression type. In actuality, it is defined exclusively for the linear based ones, and with a slight extension for polynomial. (unless I am remembering wrong from when I was implementing things...)

Anyway, shouldn't be showing those funny looking results there. Thanks!