| Author |
 Topic  |
|
|
hancok
Norway
Posts |
 Posted - 11/25/2004 : 3:28:14 PM
|
Origin Version (Select Help-->About Origin): 7.0
Operating System: Windows NT 2000
Hi,
I am trying to integrate (non graphically) the function by programming:
F(x)= x^2 exp (-1/x) from x=0.1 to 3.5
I have found the command "integrate" but when I write the following lines nothing happens...:
for (i=1; i<=31; i ++)
{
data1_C[i]= data1_B[i]^2*exp(-1/data1_B[i]);
};
integrate -r data1_C data1_B;
I would like to have the result of the integration in the column data1_D for example.
Is it possible?
thanks a lot for your helps.
Best regards,
Marc |
|
|
Mike Buess
USA
3037 Posts |
Posted - 11/25/2004 : 5:03:12 PM
|
Hi Marc,
Use this after the integate command...
Data1_D=_integ_Area;
Mike Buess
Origin WebRing Member |
 |
|
|
hancok
Norway
Posts |
Posted - 11/26/2004 : 04:23:45 AM
|
Thanks Mike!
|
|
|
|
hancok
Norway
Posts |
Posted - 11/26/2004 : 06:52:15 AM
|
Hi again,
it seems I have now a strange result of the integrale:
Data1_a(x)
1
6
11
16
21
26
31
36
41
46
51
56
61
66
71
76
81
86
91
96
101
Data1_b(y) which is the result of the function I want to integrate.
5,97531E-56
2,15E-12
4,75426E-8
2,71862E-6
2,63356E-5
1,16839E-4
3,41206E-4
7,73934E-4
0,00149
0,00255
0,00401
0,00592
0,00833
0,01126
0,01475
0,01882
0,02349
0,02878
0,0347
0,04127
0,0485
if I write the commands:
integ data1_b;
if (mks1 >= 0)
{
i1 = mks1;
i2 = mks2;
}
else
{
get data1_a -b i1;
get data1_a -e i2;
}
type -a "Integ of data1_b from zero";
type "i = $(i1) --> $(i2)";
type "x = $(integ.x1) --> $(integ.x2)";
type "Area\tPeak at\tWidth\tHeight";
type "$(integ.area)\t$(integ.x0)\t$(integ.dx)\t$(integ.y0)";
I get in a script window:
Integ of data1_b from zero
i = 1 --> 21
x = 1 --> 101
Area Peak at Width Height
1,10432 101 15 0,0485
But if I write the commands:
integrate -r Data1_b Data1_a;
Data1_c=_integ_area;
then I get:
Data1_C(y)
0
3,5
12
25,5
44
67,5
96
129,5
168
211,5
260
313,5
372
435,5
504
577,49999
655,99999
739,49998
827,99996
921,49994
1019,99992
Here, the values are way too high. I guess each value in data1_c that I get should represent the area of a small trapeze(or rectangle) and the summ of these values should give the total aera under the curve data1_b(y) which is 1,10432 according the first calculation.
However, if I add all the values from data1_c I will get a much bigger number than 1,10432.
So I guess I am doing something wrong somewhere because to me these two ways of programming should give the same results....?
Hope you guys can help me..
thanks a lot.
Marc |
|
|
|
Mike Buess
USA
3037 Posts |
Posted - 11/26/2004 : 07:57:56 AM
|
Hi Marc,
Sorry, I'm a bit confused. Why are you using the X column as an integration baseline? Also, the integral is the final value of Data1_C, or integ.area, not the sum of Data1_C.
Mike Buess
Origin WebRing Member |
|
|
|
hancok
Norway
Posts |
Posted - 11/26/2004 : 08:38:28 AM
|
Actually, I thought the integration baseline was the values over which I integrate the function
f(x)=x^2*exp(-1/x).
Since I want to integrate f(x) from x=1(=Data1_a[1]) to x=101(=data1_a[21]) I thought the X column should be the integration baseline...
The column Data1_b is the result of f(x) for the values x = data1_a .
So I guess I misunderstood the used of the intergration baseline...
So what values for the integration baseline should I used then?
Marc |
|
|
|
hancok
Norway
Posts |
Posted - 11/26/2004 : 09:07:36 AM
|
OK
I just wrote:
integrate data1_b;
and it worked...
So that was the intergration baseline I put which was my problem...
thanks !
Marc |
|
|
|
Mike Buess
USA
3037 Posts |
Posted - 11/26/2004 : 09:52:22 AM
|
Hi Marc,
The integration baseline is a dataset that needs to be subtracted (prior to integration) from the dataset you want to integrate. Typically it represents an instrumental artifact such as a DC, sloping or rolling curve riding on the "true" curve. The integration limits are taken directly from the X column... from x1=Data1_a[1] to xN=Data1_a[N].
...You can restrict the integration region to a subset of the full range, say rows 8 through 16, like this...
integrate data1_b -b 8 -e 16;
Mike Buess
Origin WebRing Member
Edited by - Mike Buess on 11/26/2004 09:57:53 AM |
|
|
|
Dizzy
Germany
Posts |
Posted - 11/30/2004 : 12:02:59 PM
|
Hi,
I would like to know how to get the area of an integrated data (value of the complete area of selected data to integrate) into a column like parameters from fitting with gauss or similar, like it is in the folowing script:
for (test=2; test<=%(data2,@#); test++){
%A=data2!wks.col$(test).label$;
Result2_a[test-1]="%A";
nlsf.begin();
nlsf.func$="gauss";
nlsf.setDepend(y,%(data2,test));
nlsf.x$="data2_A";
nlsf.constr$="";
nlsf.p1=100;
nlsf.p2=3.7;
nlsf.p3=0.02;
nlsf.p4=780;
nlsf.tolerance=0;
nlsf.fit(15);
Result2_b[test-1]=nlsf.p4;
Result2_c[test-1]=nlsf.p3;
nlsf.end();
}
Here I get a column with the values of the desired parameter. But with a loop above and integrate command I have no idea to manage that. The problem is to get only one value of integration not the whole function. When I try to use data1_b=integ.area I get the same value listed through out the entire column.
Any simple things to do for solving the problem?
Thanks - this forum helps me a lot!!
Tom |
|
|
|
Mike Buess
USA
3037 Posts |
Posted - 11/30/2004 : 2:11:14 PM
|
If you want to put the area in separate column, say Result2_d[test-1], you can do this...
for (test=2; test<=%(data2,@#); test++){
%A=data2!wks.col$(test).label$;
Result2_a[test-1]="%A";
nlsf.begin();
nlsf.func$="gauss";
//nlsf.setDepend(y,%(data2,test));
nlsf.fitdata$=%(data2,test); // I recommend this... simpler than above
nlsf.x$="data2_A";
nlsf.constr$="";
nlsf.p1=100;
nlsf.p2=3.7;
nlsf.p3=0.02;
nlsf.p4=780;
nlsf.tolerance=0;
nlsf.fit(15);
Result2_b[test-1]=nlsf.p4;
Result2_c[test-1]=nlsf.p3;
nlsf.end();
integ -a -q %(data2,test); // integrate without shading or creating _integ_area dataset
Result2_d[test-1]=integ.area; // integral area
}
Of course you need to create col D in the Result2 wks before you try this.
Mike Buess
Origin WebRing Member
Edited by - Mike Buess on 11/30/2004 2:13:44 PM |
|
|
|
Dizzy
Germany
Posts |
Posted - 12/01/2004 : 07:47:59 AM
|
Hi Mike!
Thanks for helping me - it works fine now.
Have a lovely advent season...
Tom
tomgm@hotmail.com |
|
|
| |
Topic |
|
|
|
Progam an integrale fonction.