Developing Quality Metadata

41

Generating Candidate Keywords

This is useful for locating misspelled words in documents or creating indexes and keywords for adding to a metadata block associated with a document in a media asset manager.

Given an arbitrary block of text in a file, this UNIX script will transform the text file into a sorted list of unique words. A list of stop-words is maintained in a separate file that is applied to the output list to remove unwanted keys.

If the stop list contains words like “the,” “and,” “of,” etc., then the result is a candidate list of keywords for indexing.

If the stop list contains all the words from a previously spell-checked document, ...