use strict;
use warnings;

my %wordData;
my $previousWord;

while (<DATA>) {
    chomp;
    my @parts = split;

    for my $part (@parts) {
        my ($word, $punct) = (lc $part) =~ /(\w+)(.*)/;

        if ($word) {
            $wordData{$previousWord}{$word}++ if $previousWord;
            $wordData{$word}{'!count'}++;
        }

        if ($word) {
            $previousWord = $word;
        } elsif ($punct) {
            $previousWord = undef;
        }
    }
}

for my $word (sort keys %wordData) {
    next if $wordData{$word}{'!count'} < 2;

    for my $secWord (sort keys %{$wordData{$word}}) {
        next if $secWord eq '!count';

        printf "'%s' is followed by '%s' %.0f%% of the time\n",
            $word,
            $secWord,
            100.0 * $wordData{$word}{$secWord} / $wordData{$word}{'!count'};
    }
}


__DATA__
Thank you all for the answers... all above for the answering. I am
trying something on bigram techniques. where i have stored bigram
count in one hash and word count in other.

What i am trying to do is: From looping those hashes at once, I want
to take the value of bigram count and divide with the value of word
count.

In case of ordering i used Tie::IxHash module from CPAN. Any better solution to
the way i approached my work will be appreciated... Thanks.