Re: How to extract the best match from matrix

[kevbot]

Hi jnarayan81,

As pme and choroba mentioned, you will get more/better help if you describe the objective of your nearest function.

I did not closely inspect your code, but I took a guess at what you were trying to do and came up with the following example.

#!/usr/bin/env perl

use strict;
use warnings;

use Data::Table;

# Create a Data::Table with headers (assuming data is tab-delimited)
my $dt = Data::Table::fromTSV( 'mat.txt', 1 );
# Get the number of rows in the Data::Table
my $n_rows = $dt->nofRow;

my $query = [ 37,35,59,70 ];

my $nearest_name = '';
my $min_dist;
foreach my $i (0..$n_rows - 1){
    my $row_ref = $dt->rowRef($i); # Get row of Data::Table as an ARRA
+Y REF
    my $name = shift @{$row_ref}; # The name is in the first column
    my $dist = dist($query, $row_ref);
    $min_dist = !defined($min_dist) ? $dist
                : $dist < $min_dist ? $dist
                : $min_dist;
    $nearest_name = $dist <= $min_dist ? $name : $nearest_name;
}

print "The nearest to: ";
print join(", ", @{$query});
print " is: $nearest_name\n";

exit;

# Calculate the Euclidean distance between two vectors
sub dist {
    my ($x, $y) = @_;

    unless(ref($x) eq 'ARRAY' and ref($y) eq 'ARRAY'){
        die "Vectors must be given as array references"
    }
    unless (scalar @{$x} == scalar @{$y}) {
        die "Vectors are not of equal length";
    }

    my $sum_sq = 0;
    my $len = scalar @{$x};
    foreach my $i (0..$len - 1) {
        $sum_sq += ($x->[$i] - $y->[$i])**2;
    }
    return sqrt($sum_sq);
}

I like the Data::Table module for manipulating tabular data, but there are many other ways to load/manipulate your data. If your definition of distance or best match is not euclidean distance then modify the dist subroutine accordingly.

UPDATE: Originally the dist sub returned $sum_sq. I changed it to return the correct euclidean distance which is sqrt($sum_sq).