Trying to FIND THE BEST FUNCTION A + B + C + D = SomeResult

  • Thread starter Thread starter tmanran
  • Start date Start date
  • Tags Tags
    Function
tmanran
Messages
4
Reaction score
0
Trying to "FIND THE BEST FUNCTION" A + B + C + D = SomeResult

Thanks for reading my question.



Hello, I am writing my own computer program. I found this forum while searching Google.

I need to "Find the Best Function" given a dataset.

It looks like this EXAMPLE:

A B C D RESULT
9 8 7 7 864.3
7 7 7 7 794.3
8 7 6 6 775.4
7 6 7 7 772.7
6 7 7 7 772.7
6 6 7 7 751.1

In other words, I have the values A B C D and I know what there result should be.

Now I just need an excel equation to predict the result.

I put the excel file with the data here

http://www.mediafire.com/?itm5jgxyldz

I've never done this before.

I found this other program called FindGraph

http://www.uniphiz.com/findgraph.htm

which you can download for free there. I think it will give me what I want.

But i don't know enough about math to get it to work.

Any nudge someone can give me in the right direction would be great!

Thanks!
 
Last edited by a moderator:
Physics news on Phys.org


Hello and welcome tmanran.

Maybe you could reformulate your question a little or take a screenshot of your data set and host it via photobucket (or some other free image hosting site). Or just attach it as a PDF.

The reason I ask is since you said

A + B + C + D = SomeResult

but obviously 9+8+7+7 is not equal to 864.3

I am sure it was just a matter of wording it correctly; you probably meant that

A*something + B*something_else + C*something_else + D*something_else = 864.1

I am not sure though.

Or you could just upload the .xls file by using the manage attachments button when you reply. Though I am not sure if new users have access to this functionality. Nothing personal, I just feel weird downloading stuff from 3rd party sites :smile: I rather go directly through PF.
 


f(a)+f(b)+f(c)+f(d)+17.8

where

x f(x)

0 0
1 9,6
2 25,6
3 45,4
4 95,1
5 153,3
6 172,5
7 194,1
8 218,4
9 239,8

has a maximum error of 0.1

don't ask me where those numbers come from
 


Hi

Sorry my question is not very clear.

Basically, I need an approximate equation such that, when you plug in A B C D, you will get a result.

This equation is likely not composed of simple addition A + B... despite my bad question wording. :)

And this result will be very close to the Result column.

Are you able to see the excel file?
 


1) Is this the question: "How do I find the equation that best fits my MULTIVARIABLE needs?"
2) If so, willem might have the answer. Does anyone know HOW to get it?
 


tmanran said:
Hi

Sorry my question is not very clear.

Basically, I need an approximate equation such that, when you plug in A B C D, you will get a result.

This equation is likely not composed of simple addition A + B... despite my bad question wording. :)

And this result will be very close to the Result column.

Are you able to see the excel file?

I can certainly see the excel file.
To find the result, you look up a, b, c and d in the table I've given, add the value's of f from the table, and add 17.8
 


The Chaz said:
1) Is this the question: "How do I find the equation that best fits my MULTIVARIABLE needs?"
2) If so, willem might have the answer. Does anyone know HOW to get it?

The first I noticed, was that all permutations of a,b,c and d gave the same result.
Then if you compare.

0 0 0 4 112,9
0 0 0 3 63,2
0 0 0 2 43,4
0 0 0 1 27,4
0 0 0 0 17,8

0 0 1 4 122,5
0 0 1 3 72,8
0 0 1 2 53
0 0 1 1 37
0 0 1 0 27,4

you can see that the results in the second set are 9.6 bigger, so that it appears that the same increase in the 4th colums always gives the same increase in the result, no matter what is in the other columns.
 


Wow thanks

i thought it was logarithmic! hah

thanks a bunch!
 
Back
Top