| T O P I C R E V I E W |
| jakub.zajic@hotmail.com |
Posted - 04/09/2018 : 5:27:06 PM
Origin Ver. 9.5.1.195 SR1, Win 8.1.
I cant understand the exponential function fit for this set of data works well:
ExpDec2 exponential function fit
0 3,04
10 2,77
20 2,52
30 2,27
40 2,09
50 1,92
60 1,75
70 1,62
80 1,51
90 1,43
100 1,36
110 1,29
120 1,24
130 1,19
140 1,14
150 1,09
160 1,05
170 1,02
180 0,99
190 0,97
200 0,95
210 0,93
220 0,91
230 0,89
240 0,87
250 0,85
260 0,83
270 0,82
280 0,8
290 0,79
300 0,78
310 0,77
320 0,73
330 0,71
340 0,7
350 0,69
360 0,67
370 0,66
380 0,64
390 0,63
400 0,62
410 0,61
420 0,63
430 0,59
440 0,57
450 0,56
460 0,55
470 0,54
480 0,54
490 0,53
500 0,52
510 0,52
520 0,51
530 0,5
532,5 0,5
But this set with the same settings gives me the following error:
0 4,1
10 3,63
20 3,29
30 3,05
40 2,74
50 2,51
60 2,29
70 2,07
80 1,91
90 1,75
100 1,62
110 1,52
120 1,44
130 1,37
140 1,3
150 1,26
160 1,21
170 1,16
180 1,13
190 1,08
200 1,04
210 1,01
220 0,98
230 0,96
240 0,93
250 0,92
260 0,9
270 0,89
280 0,87
290 0,86
300 0,85
310 0,84
320 0,83
330 0,83
340 0,82
350 0,81
360 0,8
370 0,79
380 0,78
390 0,77
400 0,76
410 0,75
420 0,73
424 0,73
Fit did not converge, because mutual dependency exists between parameters. The model is over-parameterized, so the fitter cannot find a fixed parameter value. Try simplifying the function, or fixing several parameter values.
Any help pls 
|
| 4 L A T E S T R E P L I E S (Newest First) |
| cc261 |
Posted - 04/12/2018 : 11:31:12 PM
quote:
Originally posted by jakub.zajic@hotmail.com
Origin Ver. 9.5.1.195 SR1, Win 8.1.
I cant understand the exponential function fit for this set of data works well:
ExpDec2 exponential function fit
0 3,04
10 2,77
20 2,52
30 2,27
40 2,09
50 1,92
60 1,75
70 1,62
80 1,51
90 1,43
100 1,36
110 1,29
120 1,24
130 1,19
140 1,14
150 1,09
160 1,05
170 1,02
180 0,99
190 0,97
200 0,95
210 0,93
220 0,91
230 0,89
240 0,87
250 0,85
260 0,83
270 0,82
280 0,8
290 0,79
300 0,78
310 0,77
320 0,73
330 0,71
340 0,7
350 0,69
360 0,67
370 0,66
380 0,64
390 0,63
400 0,62
410 0,61
420 0,63
430 0,59
440 0,57
450 0,56
460 0,55
470 0,54
480 0,54
490 0,53
500 0,52
510 0,52
520 0,51
530 0,5
532,5 0,5
But this set with the same settings gives me the following error:
0 4,1
10 3,63
20 3,29
30 3,05
40 2,74
50 2,51
60 2,29
70 2,07
80 1,91
90 1,75
100 1,62
110 1,52
120 1,44
130 1,37
140 1,3
150 1,26
160 1,21
170 1,16
180 1,13
190 1,08
200 1,04
210 1,01
220 0,98
230 0,96
240 0,93
250 0,92
260 0,9
270 0,89
280 0,87
290 0,86
300 0,85
310 0,84
320 0,83
330 0,83
340 0,82
350 0,81
360 0,8
370 0,79
380 0,78
390 0,77
400 0,76
410 0,75
420 0,73
424 0,73
Fit did not converge, because mutual dependency exists between parameters. The model is over-parameterized, so the fitter cannot find a fixed parameter value. Try simplifying the function, or fixing several parameter values.
Any help pls
Root of Mean Square Error (RMSE): 0.018230344982159
Sum of Squared Residual: 0.0146232010394153
Correlation Coef. (R): 0.999763658271671
R-Square: 0.999527372400754
Adjusted R-Square: 0.999504317395913
Determination Coef. (DC): 0.999527372400754
Chi-Square: 0.00374073581602579
F-Statistic: 20619.5996519373
Parameter Best Estimate
---------- -------------
y0 0.831478867271518
a1 3.26175955962088
t1 73.1388028122156
a2 -0.00125649117298763
t2 -94.4589349694736 |
| yuki_wu |
Posted - 04/10/2018 : 10:57:54 PM
Hi,
You choose ExpDec2: y = y0 + A1 * exp(-x/t1) +A2 * exp(-x/t2)
In fact we could rewrite the function as:
y = y0 + (B1 * exp(-x)) ^ t1 + (B2 * exp(-x)) ^ t2
A1 = B1^t1, A2 = B2^t2
If the value of B2 and t2 change but y remains the same during the fitting procedure, the fitting function will infinite solutions and the fitting doesn’t converge. This depends on whether the parameters have the same differential with respect to independent variables, not the same function you choose or how similar the datasets are.
For your case, I think you could
1. Fix the value of A2
2. Select ExpDec1 as the fitting function
3. Apply constraints to the fit process based on your experiment
Regards,
Yuki
OriginLab
|
| jakub.zajic@hotmail.com |
Posted - 04/10/2018 : 1:41:37 PM
quote:
Originally posted by yuki_wu
Hi,
Please take a look at the introduction about Over-parameterized functions, I think it should help:
https://www.originlab.com/doc/Origin-Help/The_Reason_Why_Fail_to_Converge#Over-parameterized_functions
Regards，
Yuki
OriginLab
Thanks for the link. I have tried to alter the original data a little, also tried to change some parameters of the fitting, but i still have he same problem. The confusing thing is, that i have few similar data sets, but fitting did not converge for this one only  |
| yuki_wu |
Posted - 04/09/2018 : 10:42:28 PM
Hi,
Please take a look at the introduction about Over-parameterized functions, I think it should help:
https://www.originlab.com/doc/Origin-Help/The_Reason_Why_Fail_to_Converge#Over-parameterized_functions
Regards，
Yuki
OriginLab
|
|
|
