| Author |
Comment/Response |
Jim
|
 05/15/08 07:43am
Hello
I have the following data set:
{{1.25893, 6.60693*10^11}, {1.58489, 4.36516*10^11}, {1.99526,
2.88403*10^11}, {2.51189, 1.90546*10^11}, {3.16228,
1.25893*10^11}, {3.98107, 8.31764*10^10}, {5.01187,
5.49541*10^10}, {6.30957, 3.63078*10^10}, {7.94328,
2.39883*10^10}, {10., 1.58489*10^10}, {12.5893,
1.04713*10^10}, {15.8489, 6.91831*10^9}, {19.9526,
4.57088*10^9}, {25.1189, 3.01995*10^9}, {31.6228,
1.99526*10^9}, {39.8107, 1.31826*10^9}, {50.1187,
8.70964*10^8}, {63.0957, 5.7544*10^8}, {79.4328,
3.80189*10^8}, {100., 2.51189*10^8}, {125.893,
1.65959*10^8}, {158.489, 1.09648*10^8}, {199.526,
7.24436*10^7}, {251.189, 4.7863*10^7}, {316.228,
3.16228*10^7}, {398.107, 2.0893*10^7}, {501.187,
1.38038*10^7}, {630.957, 9.12011*10^6}, {794.328,
6.0256*10^6}, {1000., 3.98107*10^6}, {1258.93,
2.63027*10^6}, {1584.89, 1.7378*10^6}, {1995.26,
1.14815*10^6}, {2511.89, 758578.}, {3162.28, 501187.}, {3981.07,
331131.}, {5011.87, 218776.}, {6309.57, 144544.}, {7943.28,
95499.3}, {10000., 63095.7}, {12589.3, 41686.9}, {15848.9,
27542.3}, {19952.6, 18197.}, {25118.9, 7621.84}, {31622.8,
3177.31}, {39810.7, 1324.52}, {50118.7, 552.154}, {63095.7,
230.176}, {79432.8, 95.9533}, {100000., 40.}, {125893.,
16.6748}, {158489., 6.9512}, {199526., 2.89774}, {251189.,
1.20798}, {316228., 0.50357}, {398107., 0.209923}}
(better to view it with a ListLogLogPlot (for mathematica 6) or LogLogListPlot (for older versions))
and I want to fit a broken power law of the form:
n[\[Gamma]_, Q1_, a1_, a2_, \[Gamma]br_] :=
Piecewise[{{Q1 \[Gamma]^-a1,
10^0.1 < \[Gamma] < \[Gamma]br}, {Q1 \[Gamma]br^(
a2 - a1) \[Gamma]^-a2, \[Gamma]br <= \[Gamma] < 10^5.6}}]
in order to get the values of Q1,a1,a2 and \[Gamma]br
The NonlinearRegress command seems that does not like the \[Gamma]br parameter. If I fix the value of \[Gamma]br to 2*10^4 then the NonlinearRegress command works perfectly. What do I have to do in order to include the \[Gamma]br into my fit?
Thanks a lot
URL: , |
|
