HP 50g: Help with quadratic regression
Stephen...@gmail.com
new HP 50g calculator. I greatly enjoy rpn. However to my horror, the
calc does not appear to be capable of preforming quadratic regression.
Now my teacher is saying "I told you so" and is making me use one of
her ti 83's. I can't stand it after having gotten used to rpn. Does
anyone know of a way to make it do quadratic regression?
scott
If you mean a non-linear (higher order) regression I suggest you try
RAS 3.0b from hpcalc.org. Here is the link
http://www.hpcalc.org/search.php?query=RAS+regression, You can
profitably spend some time looking through the great list of programs
available at hpcalc.org.
Documentation for RAS is ok. It works well for me. There are other
non-linear regression programs, but this one also gives some useful
parameters that I did not find elsewhere.
The HP is definitely better than an TI-83. I am a college physics
instructor and my courses support both TI and HP.
Good luck with your new toy! Hope you find it to be great, I do.
Regards,
Scott
John H Meyers
> I have just gone against the will of all of my teachers and bought
> a new HP 50g calculator. I greatly enjoy rpn. However to my horror,
> the calc does not appear to be capable of preforming quadratic regression.
A programmable system is capable of anything programmable,
and over the 16-year past history of the HP48/49/50 series,
most anything commonly used by many people has been programmed:
http://www.hpcalc.org/search.php?query=pregr
The "49/50 series" version should be edited
after downloading to the calculator:
http://groups.google.com/group/comp.sys.hp48/msg/804f65a58e2c3976
Or copy the UserRPL source program (for all HP48/49/50) from:
http://groups.google.com/group/comp.sys.hp48/msg/c858942c7e7d2fbc?output=gplain
Transfer between computer and calculator may be done
via cable and software, or via an SD card.
[r->] [OFF]
A.L.
wrote:
Buy Casio that costs 15 bucks. it HAS quadratic regression plus
several other.
HP is only for these specific peopke having orgasm when they hear
"RPN". Or HP.
A.L.
TW
> several other.
>
> "RPN". Or HP.
This is one of the few times when I will agree with you. It is
completely unacceptable that the high end HPs don't have more than a
basic set of regressions. Yes I know the ROM is full in most places,
but it can obviously be done since the 39/40s have all the "expected"
regressions.
That being said, I've never used anything other than a linear
regression on data with the calculator. Anything else and its time to
get out a laptop due to amount of data or more complicated tests that
don't belong on a calculator.
To Stephen:
PREGR that was pointed out does what you'd like. The TI-83 has the
worst interface of any calculator ever. I have never met anyone that
thinks it is easy to use beyond a few basic tasks. Other TI
calculators have much better usability/capability. The absolute
easiest to use is the 39/40 HPs. They are very limited when it comes
to programming though. . .
If you have any more questions, first search the newsgroup to see if
you can find the answer. Then if you can't, post it.
TW
Wes
There are numerous such programs at hpcalc.org
http://www.hpcalc.org/search.php?query=polynomial+fit
The HP User's Guide, page 18-59 (p626 in the pdf) addresses
"Polynomial Fitting" and gives a sample program to accomplish this.
The program is okay and will likely suit your needs unless you have
LOTS of data.
> I have just gone against the will of all of my teachers
As a fellow teacher, I can understand (while at the same time,
disagree) with your teachers. I prefer the 50g, but I only recommend
it to those for which it will help, not hinder, their learning of the
concepts I'm trying to teach them. I recommend it to those that are
good at thinking outside the box and can handle the learning curve.
Riches await those (such as yourself, I presume) that fit this
description. All the best.
-wes
Virgil
<7b9e5337-1012-46ce...@b64g2000hsa.googlegroups.com>,
Stephen...@gmail.com wrote:
If you mean fitting an equation of the form
y = a*x^2 + b*x + c
to a set of data points
[ x1, y1], [x2, y2], ..., [xn, yn]
where n >= 3, by least squares, it is not difficult.
For example, assuming that the x,y values are stored as rows in a two
column sigmaDAT matrix, where 'sigma' represents the upper case sigma
symbol on the HP,consider:
\<< sigmaDAT TRN AXL EVAL AXL OVER SQ AXL ROT DUP AXL SWAP 0 * 1 ADD AXL
3 COL\-> LSQ RCLVX PEVAL EVAL DUP STEQ />>
The crux of the program is that if one has the column of y values as a
vector on the stack followed by the matrix whose columns are the squares
of the x's, the x's and a third column of all ones, then the command
LSQ will return the vector of coefficients, in order of decreasing
degree in x, of the least squares parabolic function.
John H Meyers
> PREGR that was pointed out does what you'd like.
And more -- it does polynomial regression to *any* degree
(is that also included in the $15 Casio?),
and is actually a fairly short and simple program.
> the 39/40s have all the "expected" regressions.
What was the development sequence of these models?
Although one may say that new functions could be added to old models,
the 50G's built-in function set is essentially the same as
was "fixed" in the original 49G, which first appeared in 1999.
The newer Kinpo OS in 49G+/50G/etc. has nothing to do with calculator functions,
but is basically an emulator to create a new hardware platform,
and ROM revisions have otherwise basically been limited to fixing bugs.
This is not necessarily a bad policy, because it is what
also allows for compatibility between all models of the entire series.
Any library (of books) can try to stock any set of "basic books
that any library should have," but opinions and preferences may differ
as to what that set should be. "Inter-library loan" extends the effective
stock of any one library to all books that are in the collections of all libraries,
and the huge library of instantly loadable calc software for the entire series
does the same for the entire HP48/49/50 _extensible_ series.
[r->] [OFF]
TW
> (is that also included in the $15 Casio?),
> and is actually a fairly short and simple program.
Yes. I looked at it. Quite elegant.
> Although one may say that new functions could be added to old models,
> the 50G's built-in function set is essentially the same as
> was "fixed" in the original 49G, which first appeared in 1999.
Yeah. It is just that is the ONE area where the built in tools on the
"top of the line" is surpassed by pretty much everything. :-(
TW
A.L.
>In article
><7b9e5337-1012-46ce...@b64g2000hsa.googlegroups.com>,
> Stephen...@gmail.com wrote:
>
>> I have just gone against the will of allof my teachers and bought a
>> new HP 50g calculator. I greatly enjoy rpn. However to my horror, the
>> calc does not appear to be capable of preforming quadratic regression.
>> Now my teacher is saying "I told you so" and is making me use one of
>> her ti 83's. I can't stand it after having gotten used to rpn. Does
>> anyone know of a way to make it do quadratic regression?
>
>If you mean fitting an equation of the form
>
>y = a*x^2 + b*x + c
>
>to a set of data points
>
>[ x1, y1], [x2, y2], ..., [xn, yn]
>
>where n >= 3, by least squares, it is not difficult.
>
Yep. They should be selling Turing machine. Everything can be done.
A.L.
Publicly Anonomous Use
\<< sigmaDAT TRN AXL EVAL AXL OVER SQ AXL ROT DUP AXL SWAP 0 * 1 ADD AXL
3 COL\-> LSQ RCLVX PEVAL EVAL DUP STEQ />>
Now before you reach for your Manual
(which does not come with the product, and is not sold anywhere)..
Try: An Introduction to Programming
HP Graphing Calculators
Second Edition
by Mervin E. Newton
Thiel College
http://www.thiel.edu/mathproject/ITPHPC/ProgHP.PDF
And: HP-50g Calculator Tutorial ( in 33 pages )
by Merv Newton
Thiel College
http://www.thiel.edu/MathProject/CalculatorLessons/Default.htm
Regards.