Text this: Recognition of the Script in Serbian Documents Using Frequency Occurrence and Co-Occurrence Analysis