lierdumoa: (writing)
[personal profile] lierdumoa
To all you math savvy people out there. This is for my [livejournal.com profile] undermistletoe fic.

So I know that Matlab can make graphs when you input functions. I need the name of a program that works the other way. Something linux compatible that can approximate equations for graphs plotted from lab data.

Date: 2005-12-13 10:11 pm (UTC)
permetaform: (::whoahot:: [SkaterGater])
From: [personal profile] permetaform
I was weak and downloaded Grace Under Pressure.

HOLY SHIT.

watch it. watch iiiiiiit. it will apply to your fic methinks.

Date: 2005-12-13 10:50 pm (UTC)
From: [identity profile] yarngeek.livejournal.com
No clue, but you might try asking at [community profile] linux.

Date: 2005-12-13 10:53 pm (UTC)
From: [identity profile] thistle90.livejournal.com
Well, lots of programs will do that (even Excel to some extent), but you have to give the program some idea of what kind of equation you want fitted (i.e. linear, log linear, polynomials of various orders, spline functions, unlimited numbers of non-linear models, etc.). It would be possible to write a program to fit a bunch of different equations to the same data and test to see which gave the best fit using the fewest estimated parameters (the fewer the better). Math geeks and scientists still use Fortran for lots of non pre-packaged math and statistics. SAS is a higher level statistics software package that could also do the above described, but getting SAS to automatically compare the goodness of fit of a bunch of very different types of equations might be tough. You would probably have to look at the output from different equations and decide for yourself which was the best. It *might* be easier to write a highly customized program using Fortran, especially if you already know the equations underlying the different statistical models, which Charlie surely would. I've got less experience on Matlab, but I'm sure you could use it in a similar was as I described for SAS.

Date: 2005-12-13 11:06 pm (UTC)
From: [identity profile] thistle90.livejournal.com
er....to simplify: any statistics software package fits equations to data and gives you descriptive statistics to let you know how well the equation describes the data. The most widely used statistics programming language is SAS. You can also write your own equations using any basic programming language, but Fortran is probably the most widely used for brute force mathematical programming.

Date: 2005-12-14 04:23 am (UTC)
jebbypal: (Default)
From: [personal profile] jebbypal
Maple 4 might. I think I recall it being ibm based.. sorry haven't used it since calc3 in college so i don't know the specifics.

Profile

lierdumoa: (Default)
lierdumoa

February 2024

S M T W T F S
    123
45678910
11121314151617
18192021222324
25262728 29  

Most Popular Tags

Style Credit

Expand Cut Tags

No cut tags
Page generated Jul. 31st, 2025 06:33 am
Powered by Dreamwidth Studios