cancel
Showing results for 
Search instead for 
Did you mean: 
cancel
Showing results for 
Search instead for 
Did you mean: 

classification of data

ValeryOchkov
24-Ruby I

classification of data

On 4/20/2010 10:23:40 AM, sisinasi wrote:
>i want to used mathcad for
>statistics... i have got 3
>group of data... (A B C)... i
>want to found relation degree
>with A-B, A-C, B-C... is
>Mathcad solve this problem?

May be this help You:
http://twt.mpei.ac.ru/ochkov/car/c_e.html

Val
http://twt.mpei.ac.ru/ochkov/v_ochkov.htm
20 REPLIES 20

On 4/20/2010 10:23:40 AM, sisinasi wrote:
>i want to used mathcad for
>statistics... i have got 3
>group of data... (A B C)... i
>want to found relation degree
>with A-B, A-C, B-C... is
>Mathcad solve this problem ?
______________________________

Alternately, you can attach your data.

jmG



for example this problem... i attach it..

Is this what you want?

Richard

yes this is... i am study these functions... thank you very much... Is there any function for three or four independent veriables?

On 4/22/2010 4:21:08 AM, sisinasi wrote:
> Is there any
>function for three or four
>independent veriables?

The regress function will handle more than two independent variables.

Richard

What is the "relation degree"? Please define it or point us to a web site that defines it.

Richard

for example; we have got a waste water... it inclued A-B-C.. we used anti A anti B anti C...
exhaust remains

anti A anti B anti C %A %B %C
10 10 10 70 80 90
50 40 30 50 60 85
20 30 40 60 70 80

how much solution exhaust for this system

Sorry, that means nothing to me. A language or translation problem perhaps. I don't know what you mean by anti A, anti B, etc, or where %A, %B came from, or what you mean by "solution exhaust".

Richard

i am studying mathcad... thank you for your helps... i know my language is bad.. sorry.. i prepare a mathcad page for my question... thank you for your interest...

On 4/22/2010 5:15:01 AM, sisinasi wrote:
>i am studying mathcad... thank
>you for your helps... i know
>my language is bad.. sorry.. i
>prepare a mathcad page for my
>question... thank you for your
>interest...
_______________________________

Same as Richard.

Your "Fuel", ..."Save as " Mathcad 11 or lower,
then lot more collabs will look at it.

jmG



thank you for your interest... i change the file..

Yes, in principle that is solvable. The best approach depends on the nature of the data. You need to fit a multidimensional surface to the data, and then find the maximum of the surface. One such surface could be a polynomial, in which case you can use regress. You could also fit an arbitrary equation using minerr. The problem with the example you have posted is that you have only 5 "measured" values. That means the surface equation cannot have more than 5 coefficients, which is not a lot when you have 4 independent variables. It is enough to fit a surface, but not enough to fit a surface that would give you a realistic answer to your problem.

Richard

On 4/22/2010 5:15:01 AM, sisinasi wrote:
>i am studying mathcad... thank
>you for your helps...
_______________________________

The "Fuel" project is missing a data set. You have the parts% for each mixture "Fuel", you have the "Pwr" but you don't have the cost ... I have added a working cost "mix". However the "mix" must be in increasing order, that may oblige to re-arrange the "Fuel" array. The sheet includes a metafile picture of what it looks like from my guesses. If you don't have the DAEP [Data Analysis Extension Pack] it will turn red because Bicubic2D is a built-in function delivered with that pack.
At least, you can see what the project would look like.

jmG

On 4/27/2010 12:56:06 AM, jmG wrote:
>On 4/22/2010 5:15:01 AM, sisinasi wrote:
>>i am studying mathcad... thank
>>you for your helps...
>_______________________________
>
>The "Fuel" project is missing a data
>set. You have the parts% for each
>mixture "Fuel", you have the "Pwr" but
>you don't have the cost ... I have added
>a working cost "mix". However the "mix"
>must be in increasing order, that may
>oblige to re-arrange the "Fuel" array.
>The sheet includes a metafile picture of
>what it looks like from my guesses. If
>you don't have the DAEP [Data Analysis
>Extension Pack] it will turn red because
>Bicubic2D is a built-in function
>delivered with that pack.
>At least, you can see what the project
>would look like.
>
>jmG



thank you very much... i study this functions...

i have got a new problem... please help me...

On 4/28/2010 8:42:57 AM, sisinasi wrote:
>i have got a new problem...
>please help me...
____________________________

Can't help: you have ABC only ... whatever it means. For the surface plot you need two more data set as demonstrated in my previous post. You are trying to apply methods that don't apply. For the project in case, you will surely find some chemical balance equation [binary diagram] already done by chemist. At any point a project starts, all previous material/matters must be true.

jmG

On 4/28/2010 12:03:03 PM, jmG wrote:
>On 4/28/2010 8:42:57 AM, sisinasi wrote:
>>i have got a new problem...
>>please help me...
>____________________________
>
>Can't help: you have ABC only ...
>whatever it means. For the surface plot
>you need two more data set as
>demonstrated in my previous post. You
>are trying to apply methods that don't
>apply. For the project in case, you will
>surely find some chemical balance
>equation [binary diagram] already done
>by chemist. At any point a project
>starts, all previous material/matters
>must be true.
>
>jmG

thank you very much

i have got a new problem.. please help me..

On 5/11/2010 4:03:53 AM, sisinasi wrote:
>i have got a new problem...please help me.
___________________

Can't help that more than the other one before . You don't have a matrix of values for each experiment. The multivariate fit is only like a sheet of paper either flat, bent or twisted. A sheet of paper does not make balloon.

jmG

thank you for your helps.. but i am beginner in mathcad... its too complex for me... thank you..
Announcements
Check out the latest
Mathcad Tip
"PTC Mathcad 15 / Prime 1-6 Update."