# test the finnverb data set and check how many items
# were correctly classified. With default settings,
# it should get 161 correct out of 173.
use strict;
use warnings;
use Algorithm::AM::Batch;
use Test::More 0.88;
plan tests => 2;
use Test::NoWarnings;
use FindBin qw($Bin);
use Path::Tiny;
my $finnverb_data = dataset_from_file(
path => path($Bin, 'data', 'finnverb.txt'),
format => 'nocommas',
);
my $count = 0;
my $batch = Algorithm::AM::Batch->new(
training_set => $finnverb_data,
end_test_hook => sub {
my ($am, $test_item, $result) = @_;
++$count if $result->result ne 'incorrect';
}
);
$batch->classify_all($finnverb_data);
is($count, 161, '161 out of 173 items correctly predicted');