Find top unigram & bigram using bash

ubigram.sh

#!/bin/bash

# cat artikel.txt | ./ubigram.sh
# results sorted by frequences
# freq, keyword
 
cat "$@" |
tr -cs "a-zA-Z0-9" '\012' | tr '[:upper:]' '[:lower:]' |
{
old="aaa."
while read new
do
    case "$old" in
    *.) : OK;;
    *)  echo "$old $new";
    echo "$old";
    esac
    old="$new"
done
} | sort | uniq -c | sort -nr -k 2 | awk '{print $1"\t"$2" "$3}'