To illustrate this process, Compound Words
Additional processing is performed on compound words (words containing hyphens or apostrophes) to allow each portion of the compound to be examined as a candidate for keyword indexing. (Only the first four parts of a compound word will be considered; any additional parts will not be considered for indexing in whole or in part.) Compound words are processed for potential indexing according to the following steps:1. The punctuation is removed, and the entire compound is considered as a candidate keyword according to the validation procedures described in the previous section. Example: “by-product” would be examined as “byproduct”.2. Then, each individual part of the compound is considered separately as a candidate keyword: Hence, “by” and “product”.3. If there is more than one punctuation character (as in “seven-year-itch”), then there will be three or more parts to deal with; these parts will be considered in various groupings by first combining each sequential pair, and then each sequential triplet, and so on.
“seven”, “year”, and “itch” [from step 2]
“sevenyear” and “yearitch” [from step 3]