AI-Perceptron-Simple
view release on metacpan or search on metacpan
lib/AI/Perceptron/Simple.pm view on Meta::CPAN
All the subroutines from C<DATA PROCESSING RELATED SUBROUTINES>, C<NERVE DATA RELATED SUBROUTINES> and C<NERVE PORTABILITY RELATED SUBROUTINES> sections are importable through tags or manually specifying them.
The tags available include the following:
=over 4
=item C<:process_data> - subroutines under C<DATA PROCESSING RELATED SUBROUTINES> section.
=item C<:local_data> - subroutines under C<NERVE DATA RELATED SUBROUTINES> section.
=item C<:portable_data> - subroutines under C<NERVE PORTABILITY RELATED SUBROUTINES> section.
=back
Most of the stuff are OO.
=cut
use Exporter qw( import );
our @EXPORT_OK = qw(
shuffle_data shuffle_stimuli
preserve save_perceptron revive load_perceptron
preserve_as_yaml save_perceptron_yaml revive_from_yaml load_perceptron_yaml
);
our %EXPORT_TAGS = (
process_data => [ qw( shuffle_data shuffle_stimuli ) ],
local_data => [ qw( preserve save_perceptron revive load_perceptron ) ],
portable_data => [ qw( preserve_as_yaml save_perceptron_yaml revive_from_yaml load_perceptron_yaml ) ],
);
=head1 DESCRIPTION
This module provides methods to build, train, validate and test a perceptron. It can also save the data of the perceptron for future use for any actual AI programs.
This module is also aimed to help newbies grasp hold of the concept of perceptron, training, validation and testing as much as possible. Hence, all the methods and subroutines in this module are decoupled as much as possible so that the actual script...
The implementation here is super basic as it only takes in input of the dendrites and calculate the output. If the output is higher than the threshold, the final result (category) will
be 1 aka perceptron is activated. If not, then the result will be 0 (not activated).
Depending on how you view or categorize the final result, the perceptron will fine tune itself (aka train) based on the learning rate until the desired result is met. Everything from
here on is all mathematics and numbers which only makes sense to the computer and not humans anymore.
Whenever the perceptron fine tunes itself, it will increase/decrease all the dendrites that is significant (attributes labelled 1) for each input. This means that even when the
perceptron successfully fine tunes itself to suite all the data in your file for the first round, the perceptron might still get some of the things wrong for the next round of training.
Therefore, the perceptron should be trained for as many rounds as possible. The more "confusion" the perceptron is able to correctly handle, the more "mature" the perceptron is.
No one defines how "mature" it is except the programmer himself/herself :)
=head1 CONVENTIONS USED
Please take note that not all subroutines/method must be used to make things work. All the subroutines and methods are listed out for the sake of writing the documentation.
Private methods/subroutines are prefixed with C<_> or C<&_> and they aren't meant to be called directly. You can if you want to. There are quite a number of them to be honest, just ignore them if you happen to see them :)
Synonyms are placed before the actual ie. technical subroutines/methods. You will see C<...> as the parameters if they are synonyms. Move to the next subroutine/method until you find something like C<\%options> as the parameter or anything that isn't...
=head1 DATASET STRUCTURE
I<This module can only process CSV files.>
Any field ie columns that will be used for processing must be binary ie. C<0> or C<1> only. Your dataset can contain other columns with non-binary data as long as they are not one of the dendrites.
There are soem sample dataset which can be found in the C<t> directory. The original dataset can also be found in C<docs/book_list.csv>. The files can also be found L<here|https://github.com/Ellednera/AI-Perceptron-Simple>.
=head1 PERCEPTRON DATA
The perceptron/neuron data is stored using the C<Storable> module.
See C<Portability of Nerve Data> section below for more info on some known issues.
=head1 DATA PROCESSING RELATED SUBROUTINES
These subroutines can be imported using the tag C<:process_data>.
These subroutines should be called in the procedural way.
=head2 shuffle_stimuli ( ... )
The parameters and usage are the same as C<shuffled_data>. See the next two subroutines.
=head2 shuffle_data ( $original_data => $shuffled_1, $shuffled_2, ... )
=head2 shuffle_data ( ORIGINAL_DATA, $shuffled_1, $shuffled_2, ... )
Shuffles C<$original_data> or C<ORIGINAL_DATA> and saves them to other files.
=cut
sub shuffle_stimuli {
shuffle_data( @_ );
}
sub shuffle_data {
my $stimuli = shift or croak "Please specify the original file name";
my @shuffled_stimuli_names = @_
or croak "Please specify the output files for the shuffled data";
my @aoa;
for ( @shuffled_stimuli_names ) {
# copied from _real_validate_or_test
# open for shuffling
my $aoa = csv (in => $stimuli, encoding => ":encoding(utf-8)");
my $attrib_array_ref = shift @$aoa; # 'remove' the header, it's annoying :)
@aoa = shuffle( @$aoa ); # this can only process actual array
unshift @aoa, $attrib_array_ref; # put back the headers before saving file
csv( in => \@aoa, out => $_, encoding => ":encoding(utf-8)" )
and
print "Saved shuffled data into ", basename($_), "!\n";
}
}
=head1 CREATION RELATED SUBROUTINES/METHODS
=head2 new ( \%options )
Creates a brand new perceptron and initializes the value of each attribute / dendrite aka. weight. Think of it as the thickness or plasticity of the dendrites.
For C<%options>, the followings are needed unless mentioned:
=over 4
lib/AI/Perceptron/Simple.pm view on Meta::CPAN
This is the passing rate to determine the neuron output (C<0> or C<1>).
Generally speaking, this value is usually between C<0> and C<1>. However, it all depend on your combination of numbers for the other options.
=back
=cut
sub new {
my $class = shift;
my $data_ref = shift;
my %data = %{ $data_ref };
# check keys
$data{ learning_rate } = LEARNING_RATE if not exists $data{ learning_rate };
$data{ threshold } = THRESHOLD if not exists $data{ threshold };
#####
# don't pack this key checking process into a subroutine for now
# this is also used in &_real_validate_or_test
my @missing_keys;
for ( qw( initial_value attribs ) ) {
push @missing_keys, $_ unless exists $data{ $_ };
}
croak "Missing keys: @missing_keys" if @missing_keys;
#####
# continue to process the rest of the data
my %attributes;
for ( @{ $data{ attribs } } ) {
$attributes{ $_ } = $data{ initial_value };
}
my %processed_data = (
learning_rate => $data{ learning_rate },
threshold => $data{ threshold },
attributes_hash_ref => \%attributes,
);
bless \%processed_data, $class;
}
=head2 get_attributes
Obtains a hash of all the attributes of the perceptron
=cut
sub get_attributes {
my $self = shift;
%{ $self->{attributes_hash_ref} };
}
=head2 learning_rate ( $value )
=head2 learning_rate
If C<$value> is given, sets the learning rate to C<$value>. If not, then it returns the learning rate.
=cut
sub learning_rate {
my $self = shift;
if ( @_ ) {
$self->{learning_rate} = shift;
} else {
$self->{learning_rate}
}
}
=head2 threshold ( $value )
=head2 threshold
If C<$value> is given, sets the threshold / passing rate to C<$value>. If not, then it returns the passing rate.
=cut
sub threshold {
my $self = shift;
if ( @_ ) {
$self->{ threshold } = shift;
} else {
$self->{ threshold };
}
}
=head1 TRAINING RELATED SUBROUTINES/METHODS
All the training methods here have the same parameters as the two actual C<train> method and they all do the same stuff. They are also used in the same way.
=head2 tame ( ... )
=head2 exercise ( ... )
=head2 train ( $stimuli_train_csv, $expected_output_header, $save_nerve_to_file )
=head2 train ( $stimuli_train_csv, $expected_output_header, $save_nerve_to_file, $display_stats, $identifier )
Trains the perceptron.
C<$stimuli_train_csv> is the set of data / input (in CSV format) to train the perceptron while C<$save_nerve_to_file> is
the filename that will be generate each time the perceptron finishes the training process. This data file is the data of the C<AI::Perceptron::Simple>
object and it is used in the C<validate> method.
C<$expected_output_header> is the header name of the columns in the csv file with the actual category or the exepcted values. This is used to determine to tune the nerve up or down. This value should only be 0 or 1 for the sake of simplicity.
C<$display_stats> is B<optional> and the default is 0. It will display more output about the tuning process. It will show the followings:
=over 4
=item tuning status
Indicates the nerve was tuned up, down or no tuning needed
=item old sum
The original sum of all C<weightage * input> or C<dendrite_size * binary_input>
=item threshold
The threshold of the nerve
=item new sum
The new sum of all C<weightage * input> after fine-tuning the nerve
=back
If C<$display_stats> is specified ie. set to C<1>, then you B<MUST> specify the C<$identifier>. C<$identifier> is the column / header name that is used to identify a specific row of data in C<$stimuli_train_csv>.
=cut
sub tame {
train( @_ );
}
sub exercise {
train( @_ );
}
sub train {
my $self = shift;
my( $stimuli_train_csv, $expected_output_header, $save_nerve_to_file, $display_stats, $identifier ) = @_;
$display_stats = 0 if not defined $display_stats;
if ( $display_stats and not defined $identifier ) {
croak "Please specifiy a string for \$identifier if you are trying to display stats";
}
# CSV processing is all according to the documentation of Text::CSV
open my $data_fh, "<:encoding(UTF-8)", $stimuli_train_csv
or croak "Can't open $stimuli_train_csv: $!";
my $csv = Text::CSV->new( {auto_diag => 1, binary => 1} );
my $attrib = $csv->getline($data_fh);
$csv->column_names( $attrib );
# individual row
ROW: while ( my $row = $csv->getline_hr($data_fh) ) {
# print $row->{book_name}, " -> ";
# print $row->{$expected_output_header} ? "æ„æž—\n" : "é…丽优å“\n";
# calculate the output and fine tune parameters if necessary
while (1) {
my $output = _calculate_output( $self, $row );
#print "Sum = ", $output, "\n";
# $expected_output_header to be checked together over here
# if output >= threshold
# then category/result aka output is considered 1
# else output considered 0
# output expected/actual tuning
# 0 0 -
# 1 0 down
# 0 1 up
# 1 1 -
if ( ($output >= $self->threshold) and ( $row->{$expected_output_header} eq 0 ) ) {
_tune( $self, $row, TUNE_DOWN );
if ( $display_stats ) {
print $row->{$identifier}, "\n";
print " -> TUNED DOWN";
print " Old sum = ", $output;
print " Threshold = ", $self->threshold;
print " New Sum = ", _calculate_output( $self, $row ), "\n";
}
lib/AI/Perceptron/Simple.pm view on Meta::CPAN
Fine tunes the nerve. This will directly alter the attributes values in C<$self> according to the attributes / dendrites specified in C<new>.
The C<%stimuli_hash> here is the same as the one in the C<_calculate_output> method.
C<%stimuli_hash> will be used to determine which dendrite in C<$self> needs to be fine-tuned. As long as the value of any key in C<%stimuli_hash> returns true (1) then that dendrite in C<$self> will be tuned.
Tuning up or down depends on C<$tune_up_or_down> specifed by the C<train> method. The following constants can be used for C<$tune_up_or_down>:
=over 4
=item TUNE_UP
Value is C<1>
=item TUNE_DOWN
Value is C<0>
=back
This subroutine should be called in the procedural way for now.
=cut
sub _tune {
my $self = shift;
my ( $stimuli_hash_ref, $tuning_status ) = @_;
my %dendrites = $self->get_attributes;
for ( keys %dendrites ) {
if ( $tuning_status == TUNE_DOWN ) {
if ( $stimuli_hash_ref->{ $_ } ) { # must check this one, it must be 1 before we can alter the actual dendrite size in the nerve :)
$self->{ attributes_hash_ref }{ $_ } -= $self->learning_rate;
}
#print $_, ": ", $self->{ attributes_hash_ref }{ $_ }, "\n";
} elsif ( $tuning_status == TUNE_UP ) {
if ( $stimuli_hash_ref->{ $_ } ) {
$self->{ attributes_hash_ref }{ $_ } += $self->learning_rate;
}
#print $_, ": ", $self->{ attributes_hash_ref }{ $_ }, "\n";
}
}
}
=head1 VALIDATION RELATED METHODS
All the validation methods here have the same parameters as the actual C<validate> method and they all do the same stuff. They are also used in the same way.
=head2 take_mock_exam (...)
=head2 take_lab_test (...)
=head2 validate ( \%options )
This method validates the perceptron against another set of data after it has undergone the training process.
This method calculates the output of each row of data and write the result into the predicted column. The data begin written into the new file or the original file will maintain it's sequence.
Please take note that this method will load all the data of the validation stimuli, so please split your stimuli into multiple files if possible and call this method a few more times.
For C<%options>, the followings are needed unless mentioned:
=over 4
=item stimuli_validate => $csv_file
This is the CSV file containing the validation data, make sure that it contains a column with the predicted values as it is needed in the next key mentioned: C<predicted_column_index>
=item predicted_column_index => $column_number
This is the index of the column that contains the predicted output values. C<$index> starts from C<0>.
This column will be filled with binary numbers and the full new data will be saved to the file specified in the C<results_write_to> key.
=item results_write_to => $new_csv_file
Optional.
The default behaviour will write the predicted output back into C<stimuli_validate> ie the original data. The sequence of the data will be maintained.
=back
I<*This method will call C<_real_validate_or_test> to do the actual work.>
=cut
sub take_mock_exam {
my ( $self, $data_hash_ref ) = @_;
$self->_real_validate_or_test( $data_hash_ref );
}
sub take_lab_test {
my ( $self, $data_hash_ref ) = @_;
$self->_real_validate_or_test( $data_hash_ref );
}
sub validate {
my ( $self, $data_hash_ref ) = @_;
$self->_real_validate_or_test( $data_hash_ref );
}
=head1 TESTING RELATED SUBROUTINES/METHODS
All the testing methods here have the same parameters as the actual C<test> method and they all do the same stuff. They are also used in the same way.
=head2 take_real_exam (...)
=head2 work_in_real_world (...)
=head2 test ( \%options )
This method is used to put the trained nerve to the test. You can think of it as deploying the nerve for the actual work or maybe putting the nerve into an empty brain and see how
well the brain survives :)
This method works and behaves the same way as the C<validate> method. See C<validate> for the details.
lib/AI/Perceptron/Simple.pm view on Meta::CPAN
# put back the array of headers before saving file
unshift @$aoa, $attrib_array_ref;
print "Saving data to $output_file\n";
csv( in => $aoa, out => $output_file, encoding => ":encoding(utf-8)" );
print "Done saving!\n";
}
=head2 &_fill_predicted_values ( $self, $stimuli_validate, $predicted_index, $aoa )
This is where the filling in of the predicted values takes place. Take note that the parameters naming are the same as the ones used in the C<validate> and C<test> method.
This subroutine should be called in the procedural way.
=cut
sub _fill_predicted_values {
my ( $self, $stimuli_validate, $predicted_index, $aoa ) = @_;
# CSV processing is all according to the documentation of Text::CSV
open my $data_fh, "<:encoding(UTF-8)", $stimuli_validate
or croak "Can't open $stimuli_validate: $!";
my $csv = Text::CSV->new( {auto_diag => 1, binary => 1} );
my $attrib = $csv->getline($data_fh);
$csv->column_names( $attrib );
# individual row
my $row = 0;
while ( my $data = $csv->getline_hr($data_fh) ) {
if ( _calculate_output( $self, $data ) >= $self->threshold ) {
# write 1 into aoa
$aoa->[ $row ][ $predicted_index ] = 1;
} else {
#write 0 into aoa
$aoa->[ $row ][ $predicted_index ] = 0;
}
$row++;
}
close $data_fh;
$aoa;
}
=head1 RESULTS RELATED SUBROUTINES/METHODS
This part is related to generating the confusion matrix.
=head2 get_exam_results ( ... )
The parameters and usage are the same as C<get_confusion_matrix>. See the next method.
=head2 get_confusion_matrix ( \%options )
Returns the confusion matrix in the form of a hash. The hash will contain these keys: C<true_positive>, C<true_negative>, C<false_positive>, C<false_negative>, C<accuracy>, C<sensitivity>. More stats like C<precision>, C<specificity> and C<F1_Score> ...
If you are trying to manipulate the confusion matrix hash or something, take note that all the stats are in percentage (%) in decimal (if any) except the total entries.
For C<%options>, the followings are needed unless mentioned:
=over 4
=item full_data_file => $filled_test_file
This is the CSV file filled with the predicted values.
Make sure that you don't do anything to the actual and predicted output in this file after testing the nerve. These two columns must contain binary values only!
=item actual_output_header => $actual_column_name
=item predicted_output_header => $predicted_column_name
The binary values are treated as follows:
=over 4
=item C<0> is negative
=item C<1> is positive
=back
=item more_stats => 1
Optional.
Setting it to C<1> will process more stats that are usually not so important eg. C<precision>, C<specificity> and C<F1_Score>
=back
=cut
sub get_exam_results {
my ( $self, $info ) = @_;
$self->get_confusion_matrix( $info );
}
sub get_confusion_matrix {
my ( $self, $info ) = @_;
my %c_matrix = _collect_stats( $info ); # processes total_entries, accuracy, sensitivity etc
%c_matrix;
}
=head2 &_collect_stats ( \%options )
Generates a hash of confusion matrix based on C<%options> given in the C<get_confusion_matrix> method.
=cut
( run in 0.682 second using v1.01-cache-2.11-cpan-cdf2f3d4e48 )