Timeline for Exponential fitting - isn't actually a BUG there?
Current License: CC BY-SA 3.0
6 events
| when toggle format | what | by | license | comment | |
|---|---|---|---|---|---|
| Feb 3, 2016 at 8:10 | comment | added | Simplex |
@SjoerdC.deVries, if I copy your code it works, but if I edit it fit1 = FindFit[data, a*Exp[b*x], {a, b}, x] then it does not. Does the edited version work for you?
|
|
| Feb 3, 2016 at 7:56 | comment | added | Sjoerd C. de Vries | @simplex are you sure that you have converted all numbers to exact numbers? I find it surprising that there should be a difference between our computers in this regard as the arbitrary precision routines are supposedly machine independent. | |
| Feb 3, 2016 at 7:53 | comment | added | Simplex | The "Exact numbers" example doesn't work for me unless a helpful starting value is given for b. Also on my computer, giving almost any value for WorkingPrecision works (even lower than $MachinePrecision). But "WorkingPrecision->MachinePrecision" does not work. | |
| Jan 28, 2016 at 10:02 | comment | added | Alexey Popkov | Please see my answer. | |
| Jan 28, 2016 at 3:02 | comment | added | Oleksandr R. |
It's sufficient in the last case just to add WorkingPrecision -> $MachinePrecision. It is b that needs to have the arbitrary precision, rather than the inputs, it seems. For machine-precision b it will fail even with increased WorkingPrecision.
|
|
| Jan 27, 2016 at 22:20 | history | answered | Sjoerd C. de Vries | CC BY-SA 3.0 |