forum.alglib.net

ALGLIB forum
It is currently Mon Dec 23, 2024 4:31 am

All times are UTC


Forum rules


1. This forum can be used for discussion of both ALGLIB-related and general numerical analysis questions
2. This forum is English-only - postings in other languages will be removed.



Post new topic Reply to topic  [ 3 posts ] 
Author Message
 Post subject: Function fitting problem
PostPosted: Fri Sep 07, 2012 10:52 am 
Offline

Joined: Fri Sep 07, 2012 10:18 am
Posts: 2
Hi there!

trying to use ALGLIB to fit experimental data.

I have an array of 101 elemens (e.g.)
they are in range of -10..1000

Trying to use gaussian to fit experimantal data:

y(x)=c0*exp[-(x-c1)^2/(2*c2^2)]

c - regression coefficients..

we used this example:
http://www.alglib.net/translator/man/ma ... sfit_d_nlf

But every time we get in info code 4.
According to (http://www.alglib.net/translator/man/ma ... fitresults) this means : gradient norm is no more than EpsG..

What is that mean? and how can we get estimates of the coefficients (they don't change from initial values)?
Maybe we do something wrong?

Here is the code in c#
(running on windows 7 x64, .NET 4.0)

Code:
//main code
private void DrawApproxGraph()
        {
            try
            {
                List<double> zond = GetZond(Convert.ToInt32(Left_textbox.Value),
                                            Convert.ToInt32(Right_textbox.Value));

                double[,] x = CreateXArr(Convert.ToInt32(Left_textbox.Value), Convert.ToInt32(Right_textbox.Value));
               
                double[] y = zond.ToArray();
                double[] c = new double[] { 1000, 20, 10};
                double epsf = 0;
                double epsx = 0.0001;
                int maxits = 0;
                int info;
                alglib.lsfitstate state;
                alglib.lsfitreport rep;
                double diffstep = 0.000000001;
                //
                // Fitting without weights
                //
                alglib.lsfitcreatef(x, y, c, diffstep, out state);
                alglib.lsfitsetcond(state, epsf, epsx, maxits);
                alglib.lsfitfit(state, function_cx_1_func, null, null);
                alglib.lsfitresults(state, out info, out c, out rep);
                System.Console.WriteLine("{0}", info); // EXPECTED: 2
                System.Console.WriteLine("{0}", alglib.ap.format(c, 1));
            }
            catch (Exception ex)
            {

            }
        }
//end of main

public static void function_cx_1_func(double[] c, double[] x, ref double func, object obj)
        {
            func = c[0] * System.Math.Exp(Math.Pow(x[0] - c[1], 2) / (-2 * Math.Pow(c[2], 2)));
        }


Last edited by twistfire on Fri Sep 07, 2012 12:11 pm, edited 2 times in total.

Top
 Profile  
 
 Post subject: Re: Function fitting problem
PostPosted: Fri Sep 07, 2012 11:39 am 
Offline

Joined: Fri Sep 07, 2012 10:18 am
Posts: 2
PS
data for testing

data
---
x {double[101, 1]} double[,]
[0, 0] 227.0 double
[1, 0] 228.0 double
[2, 0] 229.0 double
[3, 0] 230.0 double
[4, 0] 231.0 double
[5, 0] 232.0 double
[6, 0] 233.0 double
[7, 0] 234.0 double
[8, 0] 235.0 double
[9, 0] 236.0 double
[10, 0] 237.0 double
[11, 0] 238.0 double
[12, 0] 239.0 double
[13, 0] 240.0 double
[14, 0] 241.0 double
[15, 0] 242.0 double
[16, 0] 243.0 double
[17, 0] 244.0 double
[18, 0] 245.0 double
[19, 0] 246.0 double
[20, 0] 247.0 double
[21, 0] 248.0 double
[22, 0] 249.0 double
[23, 0] 250.0 double
[24, 0] 251.0 double
[25, 0] 252.0 double
[26, 0] 253.0 double
[27, 0] 254.0 double
[28, 0] 255.0 double
[29, 0] 256.0 double
[30, 0] 257.0 double
[31, 0] 258.0 double
[32, 0] 259.0 double
[33, 0] 260.0 double
[34, 0] 261.0 double
[35, 0] 262.0 double
[36, 0] 263.0 double
[37, 0] 264.0 double
[38, 0] 265.0 double
[39, 0] 266.0 double
[40, 0] 267.0 double
[41, 0] 268.0 double
[42, 0] 269.0 double
[43, 0] 270.0 double
[44, 0] 271.0 double
[45, 0] 272.0 double
[46, 0] 273.0 double
[47, 0] 274.0 double
[48, 0] 275.0 double
[49, 0] 276.0 double
[50, 0] 277.0 double
[51, 0] 278.0 double
[52, 0] 279.0 double
[53, 0] 280.0 double
[54, 0] 281.0 double
[55, 0] 282.0 double
[56, 0] 283.0 double
[57, 0] 284.0 double
[58, 0] 285.0 double
[59, 0] 286.0 double
[60, 0] 287.0 double
[61, 0] 288.0 double
[62, 0] 289.0 double
[63, 0] 290.0 double
[64, 0] 291.0 double
[65, 0] 292.0 double
[66, 0] 293.0 double
[67, 0] 294.0 double
[68, 0] 295.0 double
[69, 0] 296.0 double
[70, 0] 297.0 double
[71, 0] 298.0 double
[72, 0] 299.0 double
[73, 0] 300.0 double
[74, 0] 301.0 double
[75, 0] 302.0 double
[76, 0] 303.0 double
[77, 0] 304.0 double
[78, 0] 305.0 double
[79, 0] 306.0 double
[80, 0] 307.0 double
[81, 0] 308.0 double
[82, 0] 309.0 double
[83, 0] 310.0 double
[84, 0] 311.0 double
[85, 0] 312.0 double
[86, 0] 313.0 double
[87, 0] 314.0 double
[88, 0] 315.0 double
[89, 0] 316.0 double
[90, 0] 317.0 double
[91, 0] 318.0 double
[92, 0] 319.0 double
[93, 0] 320.0 double
[94, 0] 321.0 double
[95, 0] 322.0 double
[96, 0] 323.0 double
[97, 0] 324.0 double
[98, 0] 325.0 double
[99, 0] 326.0 double
[100, 0] 327.0 double
---
y {double[101]} double[]
[0] -0.17 double
[1] 0.0 double
[2] -0.17 double
[3] -0.33 double
[4] -0.17 double
[5] -0.17 double
[6] -0.17 double
[7] -0.33 double
[8] -0.17 double
[9] -0.33 double
[10] -0.33 double
[11] -0.33 double
[12] -0.17 double
[13] -0.17 double
[14] -0.17 double
[15] 0.0 double
[16] 0.0 double
[17] -0.17 double
[18] -0.17 double
[19] -0.17 double
[20] -0.17 double
[21] -0.17 double
[22] 0.16 double
[23] 0.48 double
[24] 1.29 double
[25] 2.43 double
[26] 3.4 double
[27] 3.72 double
[28] 3.72 double
[29] 3.24 double
[30] 2.26 double
[31] 1.29 double
[32] 3.72 double
[33] 14.25 double
[34] 44.06 double
[35] 101.41 double
[36] 190.5 double
[37] 310.38 double
[38] 451.8 double
[39] 600.52 double
[40] 737.89 double
[41] 850.8 double
[42] 932.29 double
[43] 982.5 double
[44] 1000.0 double
[45] 983.15 double
[46] 929.69 double
[47] 847.4 double
[48] 743.72 double
[49] 627.08 double
[50] 510.93 double
[51] 409.04 double
[52] 321.88 double
[53] 248.66 double
[54] 190.18 double
[55] 147.25 double
[56] 118.58 double
[57] 100.92 double
[58] 88.93 double
[59] 78.4 double
[60] 67.55 double
[61] 57.02 double
[62] 47.14 double
[63] 39.04 double
[64] 32.88 double
[65] 27.37 double
[66] 20.89 double
[67] 14.09 double
[68] 8.1 double
[69] 3.72 double
[70] 1.45 double
[71] 0.64 double
[72] 0.64 double
[73] 0.81 double
[74] 0.97 double
[75] 0.97 double
[76] 1.78 double
[77] 3.4 double
[78] 5.83 double
[79] 8.26 double
[80] 9.39 double
[81] 9.23 double
[82] 8.1 double
[83] 6.15 double
[84] 4.69 double
[85] 4.53 double
[86] 5.02 double
[87] 5.83 double
[88] 6.8 double
[89] 6.96 double
[90] 7.12 double
[91] 7.29 double
[92] 7.61 double
[93] 8.1 double
[94] 8.74 double
[95] 8.42 double
[96] 7.45 double
[97] 5.99 double
[98] 4.05 double
[99] 2.59 double
[100] 1.13 double
---


Top
 Profile  
 
 Post subject: Re: Function fitting problem
PostPosted: Fri Sep 07, 2012 12:43 pm 
Offline
Site Admin

Joined: Fri May 07, 2010 7:06 am
Posts: 927
I think the problem is that you've passed bad initial values. Your data clearly have peak somewhere at x=271, but you initial estimate is a Gaussian centered at x=20 with width=10. It is effectively zero at x=271, so gradient with respect to all of its parameters is zero.

Try providing better estimates - choose location of maximum x as initial location of your Gaussian, y-value at this x as coefficient before Gaussian, and choose some reasonable width (it is better to choosing it from data, not to use some fixed value). Does it solve your problem?


Top
 Profile  
 
Display posts from previous:  Sort by  
Post new topic Reply to topic  [ 3 posts ] 

All times are UTC


Who is online

Users browsing this forum: No registered users and 8 guests


You cannot post new topics in this forum
You cannot reply to topics in this forum
You cannot edit your posts in this forum
You cannot delete your posts in this forum
You cannot post attachments in this forum

Search for:
Jump to:  
cron
Powered by phpBB © 2000, 2002, 2005, 2007 phpBB Group