NAME

TimeSeries::AdaptiveFilter - Adaptive filter for data stream with possible outliers

VERSION

Version 0.04

STATUS

SYNOPSYS

use TimeSeries::AdaptiveFilter qw/filter/;

# creation with defaults
my $filter = filter();

# create filter with tuned parameters
my $filter = filter({
  floor             => 6
  cap               => 0.2,
  lookback_capacity => 20,
  lookback_period   => 4,
  decay_speeds      => [0.03, 0.01, 0.003],
  build_up_count    => 5,
  reject_criterium  => 4,
});

# usage
my $now = time;
$filter->($now, 100.002);        # returns true, i.e. all data is valid on learning period
$filter->($now + 1, 100.001);    # returns true
...                              # it learns form sample of 60 seconds
$filter->($now + 60, 100.005);   # returns true
$filter->($now + 61, 99.9995);   # returns true, as value does not differs much
$filter->($now + 62, 10_0000);   # returns false, outlier data
$filter->($now + 63, 10.0001);   # returns false, outlier data
$filter->($now + 64, 100.011);   # returns true, even if the sample is oulier, because
                                 # the filter rejected too much values, and has to
                                 # re-adapt to time seria again

DESCRIPTION

For the details of underlying mathematical model of the filter, configurable paramters and their usage, please, look at the shipped doc folder.