The Origin Forum
File Exchange
Try Origin for Free
The Origin Forum
Home | Profile | Register | Active Topics | Members | Search | FAQ | Send File to Tech support
Username:
 Password:
Save Password
Forgot your Password? | Admin Options

 All Forums
 Origin Forum
 Origin Forum
 non linear curve fit		  New Topic  Reply to Topic
 Printer Friendly
Author Previous Topic Topic Next Topic Lock Topic Edit Topic Delete Topic New Topic Reply to Topic

premranjan

India
Posts
Posted - 11/10/2007 :  2:00:20 PM  Show Profile  Edit Topic  Reply with Quote  View user's IP address  Delete Topic
Origin Version (Select Help-->About Origin): Origin pro 6.1
Operating System:xp
i am using a non linear curve fit.here are the x and y values
x y
0.056 4.48176
0.198 6.67682
0.311 6.77246
0.4299 7.15525
0.4693 7.35968
0.6999 10.06429
0.7452 10.50408
0.8174 14.21204
0.973 18.72605
i am using the equation y=(P1*P2*P3x)/(1-P2x).i have used the following steps plot-scatter-analysis-nonlinear curve fit-user defined function-equation-parametr 3-given the parameter value..but while i itereate error box of use the y script or equation form...HELP ME PLEASE

Mike Buess

USA
3037 Posts
Posted - 11/10/2007 :  9:38:55 PM  Show Profile  Edit Reply  Reply with Quote  View user's IP address  Delete Reply
You can eliminate the error message by following its instructions and setting the function form to Y-Script or Equations (either will work) as shown in the figure below. However, there are other problems with your fitting function. For one thing, two parameters (P1 and P2) are sufficient in the numerator as indicated by the dependencies of P1 and P3 which are exactly 1 after each iteration. Even after the superfluous P3 is removed from numerator the fit to your data never reaches stability. (Chi^2 changes with every iteration.) Try adding a y-offset parameter like this...

y = P3 + (P1*P2*x) / (1-P2*x)



Mike Buess
Origin WebRing Member
Go to Top of Page

premranjan

India
Posts
Posted - 11/11/2007 :  2:02:43 PM  Show Profile  Edit Reply  Reply with Quote  View user's IP address  Delete Reply
Thank u mike
i have used the equation y=P1*P2*P3*x/(1-P3*x)*(1-P3*x+P3*P2*x)
i draw the graph as u suggested using nonlinear curvefit&equation mode
But instead of getting smooth 'S'shape curve i am getting 'Z' shape after iteareate.i am taking P1 as 5 ,P2 as 4 and P3 as .44 as my constant.can u tell me where i am getting wrong???
Go to Top of Page

Mike Buess

USA
3037 Posts
Posted - 11/11/2007 :  6:02:29 PM  Show Profile  Edit Reply  Reply with Quote  View user's IP address  Delete Reply
When I fit your data with the new equation starting with P1=5, P2=4 and P3=0.44 (fixed) the fit curve approaches a straight line with P1>>1 and P2<1. My guess is that you're still using the wrong fitting equation. The fit to data is again better if you add a Y-offset.

Mike Buess
Origin WebRing Member

Edited by - Mike Buess on 11/11/2007 6:12:29 PM

Edited by - Mike Buess on 11/11/2007 6:14:12 PM
Go to Top of Page

premranjan

India
Posts
Posted - 11/12/2007 :  12:23:36 AM  Show Profile  Edit Reply  Reply with Quote  View user's IP address  Delete Reply
yah mike pls use these data and eqn is same but u have to give bracket
like y=(P1*P2*P3*x)/((1-P3*x)*(1-P3*x+P3*P2*x))and main thing is go to
action - dataset-assign X(first) -row 1-29 then assign the otherwith same row value.Take P1,P2,P3 value same .Yah one thing is while writeing eqn at function-select should be the right part be user12 or any and left part isin expontial form and see the graph coming like )
form but i need S form. /
x y
0.0962 0.041
0.0962 0.04332
0.211 0.0672
0.211 0.06988
0.3205 0.07111
0.3205 0.07221
0.4364 0.07402
0.4364 0.07605
0.4991 0.07801
0.4991 0.07833
0.7206 0.0801
0.7206 0.08113
0.7487 0.08212
0.7487 0.09042
0.8295 0.10543
0.8295 0.10522
0.9671 0.16142
0.9671 0.16121
Go to Top of Page

Mike Buess

USA
3037 Posts
Posted - 11/12/2007 :  01:48:53 AM  Show Profile  Edit Reply  Reply with Quote  View user's IP address  Delete Reply
Function now converges but the fit is still not good. I don't really understand your actions... "dataset-assign X(first) -row 1-29 then assign the otherwith same row value" ... your data only have 18 rows. Either you have insufficient data or the equation is still not right.

Mike Buess
Origin WebRing Member
Go to Top of Page

premranjan

India
Posts
Posted - 11/12/2007 :  02:17:08 AM  Show Profile  Edit Reply  Reply with Quote  View user's IP address  Delete Reply
Thanks mike

but how can i upload my graph to you...from origin data sheet which i had drawn by using the equation and which type graph i wanted..the file is not uploading but it is in .opg or .org type..actually i am doing my project on sorption isotherm and my guide also doesn't not know how to draw graph in origin.
Go to Top of Page

Mike Buess

USA
3037 Posts
Posted - 11/12/2007 :  03:48:35 AM  Show Profile  Edit Reply  Reply with Quote  View user's IP address  Delete Reply
Pasting the data in the message like you did before works fine. The following graph is what I get fitting your last equation to your latest set of data. Parameters were initialized as P1=5, P2=4 and P3 was fixed as 0.44 as in your previous post. Perhaps the fit shown below is good enough for your purposes? I did notice that a few of your X-values have more than one Y-values. You might be able to decrease the noise in your data with the methods discussed in this topic...

http://www.originlab.com/forum/topic.asp?TOPIC_ID=6170



quote:
my guide also doesn't not know how to draw graph in origin.
Just select the Y column and choose Plot> Scatter.

Mike Buess
Origin WebRing Member

Edited by - Mike Buess on 11/12/2007 04:02:59 AM
Go to Top of Page

premranjan

India
Posts
Posted - 11/12/2007 :  10:00:42 AM  Show Profile  Edit Reply  Reply with Quote  View user's IP address  Delete Reply
Thank you very much MIKE....
I am grateful to you.The only thing i am going wrong is that i am not fixing the parameter P3.when i fixed it i got the perfect graph.
Thanks \\
Go to Top of Page
  Previous Topic Topic Next Topic Lock Topic Edit Topic Delete Topic New Topic Reply to Topic
 New Topic  Reply to Topic
 Printer Friendly
Jump To:
The Origin Forum © 2020 Originlab Corporation Go To Top Of Page
Snitz Forums 2000