# Copyright 2007, 2009, 2010, 2011 Kevin Ryde # This file is part of Chart. # # Chart is free software; you can redistribute it and/or modify it under the # terms of the GNU General Public License as published by the Free Software # Foundation; either version 3, or (at your option) any later version. # # Chart is distributed in the hope that it will be useful, but WITHOUT ANY # WARRANTY; without even the implied warranty of MERCHANTABILITY or FITNESS # FOR A PARTICULAR PURPOSE. See the GNU General Public License for more # details. # # You should have received a copy of the GNU General Public License along # with Chart. If not, see . package App::Chart::Series::Derived::MedianAverage; use 5.010; use strict; use warnings; use Carp; use List::Util qw(min max); use POSIX (); use Locale::TextDomain ('App-Chart'); use base 'App::Chart::Series::Indicator'; use App::Chart::Series::Derived::EMA; use App::Chart::Series::Derived::TMA; # http://www.mesasoftware.com/technicalpapers.htm # http://www.mesasoftware.com/Papers/What's%20the%20Difference.exe # http://web.archive.org/web/20070720222047/http://www.mesasoftware.com/Papers/What%27s+the+Difference.exe # Original gone, use archive.org. # John Ehler's paper. Sample chart of something unspecified. # sub longname { __('Median-Average Adaptive') } sub shortname { __('Median-Average') } sub manual { __p('manual-node','Median-Average Adaptive Filter') } use constant { type => 'average', parameter_info => [ ], }; sub new { my ($class, $parent) = @_; return $class->SUPER::new (parent => $parent, parameters => [ ], arrays => { values => [] }, array_aliases => { }); } use constant warmup_count => (App::Chart::Series::Derived::TMA->warmup_count(4) + 39 # lookback for median + App::Chart::Series::Derived::EMA->warmup_count(39)); # slowest smoothing ### MedianAverage warmup_count(): warmup_count() sub proc { my ($class) = @_; my $proc_average_and_alpha = $class->proc_average_and_alpha; return sub { return ($proc_average_and_alpha->(@_))[0]; }; } my @alpha_array = map {App::Chart::Series::Derived::EMA::N_to_alpha($_)} (0 .. 39); use constant THRESHOLD => 0.002; sub proc_average_and_alpha { my ($class) = @_; my $smooth_proc = App::Chart::Series::Derived::TMA->proc (4); # last 39 smoothed input values my @values; # $prev is the previous median-average value calculated. $prev is # initialized to the first SMOOTHed input. Ehler's code and some # of the Trader's tips show an initial zero, which makes the filter rise # up from zero at the start of the data. Since each alpha determined # depends on the previous value there's no way to chop off an infinite # sequence like an ordinary EMA. Using the first smoothed should be # reasonable, it won't take too long for the medians to start moving and # the average tracking towards that. # my $prev; return sub { my ($value) = @_; my $alpha; my $smooth = $smooth_proc->($value); $prev //= $smooth; # initial unshift @values, $smooth; if (@values > 39) { pop @values; } my $len = round_down_odd (scalar @values); for (;;) { my @sorted = sort @values[0 .. $len-1]; my $median = $sorted[$len/2]; $alpha = $alpha_array[$len]; my $average = $smooth * $alpha + $prev * (1 - $alpha); my $ratio = ($median == 0 ? 0 : abs($median-$average) / $median); if ($ratio <= THRESHOLD || $len <= 3) { $prev = $average; return ($average, $alpha); } $len -= 2; } }; } sub round_down_odd { my ($x) = @_; return 2 * POSIX::floor (($x-1) / 2) + 1; } 1; __END__ # =head1 NAME # # App::Chart::Series::Derived::MedianAverage -- Median-Average Adaptive Filter # # =head1 SYNOPSIS # # my $series = $parent->MedianAverage(); # # =head1 DESCRIPTION # # ... # # =head1 SEE ALSO # # L # # =cut