Re^2: fitting double lognormal distribution on a dataset

G'day hdb,

... the documentation of Algorithm::CurveFit seems to specify that powers need to be written as x^n and not x**n.

Firstly, I will say that the first line of code in the Algorithm::CurveFit SYNOPSIS:

my $formula = 'c + a * x^2';
[download]

is written in a way that suggests this intent

$ perl -MO=Deparse,-p -e 'my $formula = $c + $a * $x**2;'
(my $formula = ($c + ($a * ($x ** 2))));
-e syntax OK
[download]

more than it does this one

$ perl -MO=Deparse,-p -e 'my $formula = $c + $a * $x^2;'
(my $formula = (($c + ($a * $x)) ^ 2));
-e syntax OK
[download]

However, further down they have

The formula should be a string that can be parsed by Math::Symbolic.

The doco for that module has

Warning: The operator to use for exponentiation is the normal Perl operator for exponentiation **, NOT the caret ^ which denotes exponentiation in the notation that is recognized by the Math::Symbolic parsers! The ^ operator will be interpreted as the normal binary xor.

[Update (cosmetic): 2 instances of quoting my own code — <blockquote>s removed.]

-- Ken

Comment on Re^2: fitting double lognormal distribution on a dataset Select or Download Code

Replies are listed 'Best First'.
Re^3: fitting double lognormal distribution on a dataset by hdb (Monsignor) on May 18, 2013 at 19:04 UTC
Thanks for the clarification. I find this documentation very confusing. If I see an example like `'1/2 * m * v^2'` I believe that the caret is the operator for exponentiation and I do not see any need to even continue reading. Obviously wrong here. A fully working example would be useful here to test such presumptions...	[reply] [d/l]
Re^4: fitting double lognormal distribution on a dataset by sdani (Novice) on May 19, 2013 at 03:48 UTC
On an made up dataset the program runs fine. So the proper way to raise to the power is this: `n ^ 2` not this: `n*2`; here is the sample code: #!/usr/bin/perl use strict; use warnings; use Algorithm::CurveFit; my @X_data = qw(-20 -19 -18 -17 -16 -15 -14 -13 -12 -11 -10 -9 -8 -7 - +6 -5 -4 -3 -2 -1 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14); my @Y_data = qw(1212 1095 984 879 780 687 600 519 444 375 312 255 204 +159 120 87 60 39 24 15 12 15 24 39 60 87 120 159 204 255 312 375 444 +519 600); my $variable = 'x'; my $equation = 'r + m x^2'; my @parameters = ( ['r', '100'], ['m', '1'], ); my $maxiter = '200'; my $square_residual = Algorithm::CurveFit->curve_fit( formula => $equation, # may be a Math::Symbolic tree in +stead params => \@parameters, variable => $variable, xdata => \@X_data, ydata => \@Y_data, maximum_iterations => $maxiter, ); print "$parameters[0][1] (should be 12)\t$parameters[1][1] (should be +3)\n"; [download] I'll try to do the more complex equation and I let you know how it went! Thanks.	[reply] [d/l] [select]