Finding Fourier formula to fit a given data set

Question

I am trying to find a model with a Fourier basis for this data:

data = {{1, -0.5}, {2, -15}, {3, 30}, {4, 184.25}, {5, 
    2143.75}, {6, 6234.75}, {7, 11969.75}, {8, 16940.75}, {9, 
    20484.75}, {10, 23084.25}, {11, 24577.25}, {12, 26321.75}, {13, 
    29709.25}, {14, 36357.75}, {15, 40502.25}, {16, 38244.25}, {17, 
    30486.25}, {18, 19492.75}, {19, 13318.25}, {20, 12267.25}, {21, 
    12376.25}, {22, 12375.75}, {23, 12376.25}, {24, 12376.25}};

Help is greatly appreciated thanks!

Can you post some actual code with the data to make it easier to copy and paste? You will find more easily help this way. Also, can you show us what you've tried so far and how it fails? Some will (hopefully) also post a link with a proper welcome - I would do it but I don't remember the appropriate link. — , Oct 02 '19 at 10:18
Possible duplicate of "Find a Fourier series from discrete data" . — Anton Antonov, Oct 02 '19 at 14:11
Another possible duplicate "Fourier Transform to help guess with NonLinearModelFit". — Anton Antonov, Oct 02 '19 at 14:41

score 1 · Answer 1 · edited Oct 02 '19 at 14:04

Is this what you wanted?

data = {{1, -0.5}, {2, -15}, {3, 30}, {4, 184.25}, {5, 
    2143.75}, {6, 6234.75}, {7, 11969.75}, {8, 16940.75}, {9, 
    20484.75}, {10, 23084.25}, {11, 24577.25}, {12, 26321.75}, {13, 
    29709.25}, {14, 36357.75}, {15, 40502.25}, {16, 38244.25}, {17, 
    30486.25}, {18, 19492.75}, {19, 13318.25}, {20, 12267.25}, {21, 
    12376.25}, {22, 12375.75}, {23, 12376.25}, {24, 12376.25}};

lm = LinearModelFit[data, x, x]

The answer is

FittedModel[7583.2 +732.805 x]

If yes, then you can read more about it here LinearModelFit

score 1 · Answer 2 · answered Oct 02 '19 at 14:54

Although I think that this question is a duplicate of others (for example, this and this) it seems that applying the suggestions/code from those answers is not direct or trivial.

Here is a solution which I found with QRMon but redressed it as a NonlinearModelFit one.

bFuncs = Join[Table[Sin[i x/20], {i, 1, 60, 2}], Table[Cos[i x/20], {i, 0, 20, 2}]];

bs = Array[b, Length[bFuncs]];

nlm = NonlinearModelFit[data, bs.bFuncs, bs, x]

Show[ListPlot[data], Plot[nlm[x], {x, Min[data[[All, 1]]], Max[data[[All, 1]]]}], Frame -> True]

ListPlot[nlm["FitResiduals"], Filling -> Axis]

nlm["Function"][x]

(* 5669.1 + 1336.03 Cos[x/10] - 8003.44 Cos[x/5] - 
 2004.01 Cos[(3 x)/10] + 1462.08 Cos[(2 x)/5] + 1225.55 Cos[x/2] - 
 676.923 Cos[(3 x)/5] - 396.566 Cos[(7 x)/10] + 
 1900.68 Cos[(4 x)/5] + 164.913 Cos[(9 x)/10] - 1037.4 Cos[x] + 
 9867.23 Sin[x/20] + 8288.73 Sin[(3 x)/20] - 2577.54 Sin[x/4] - 
 3047.06 Sin[(7 x)/20] - 540.337 Sin[(9 x)/20] + 
 887.6 Sin[(11 x)/20] - 403.283 Sin[(13 x)/20] - 
 1435.52 Sin[(3 x)/4] + 1099.31 Sin[(17 x)/20] + 
 981.363 Sin[(19 x)/20] - 556.755 Sin[(21 x)/20] - 
 132.728 Sin[(23 x)/20] + 103.919 Sin[(5 x)/4] + 
 546.033 Sin[(27 x)/20] - 243.036 Sin[(29 x)/20] - 
 441.13 Sin[(31 x)/20] + 389.737 Sin[(33 x)/20] - 
 297.45 Sin[(7 x)/4] + 96.2562 Sin[(37 x)/20] + 
 285.294 Sin[(39 x)/20] - 343.174 Sin[(41 x)/20] + 
 194.906 Sin[(43 x)/20] - 74.1913 Sin[(9 x)/4] + 
 37.3024 Sin[(47 x)/20] + 108.943 Sin[(49 x)/20] - 
 308.276 Sin[(51 x)/20] + 337.058 Sin[(53 x)/20] - 
 127.753 Sin[(11 x)/4] + 87.3725 Sin[(57 x)/20] - 
 29.0636 Sin[(59 x)/20] *)

Finding Fourier formula to fit a given data set

2 Answers2