Which list cleanup (.fuz) file do I use for which fields?
• General – the general.fuz file is the most strict in its matching. The general file will only match differences in capitalisation, punctuation and stemming. It is best used with descriptor fields and NLP terms to catch inconsistencies. It is very accurate, but the stemming will sometimes match more technical terms than you want (magnets and magnetism for example), so you may want to check its conclusions before accepting them. • Author – the author.fuz file is strict in matching words, but is less concerned about the order they are in. It will also allow small variations to account for initials. For example, it will match “John B Smith” to “Smith, J B”. As you would expect, it is best used on inventor fields. • Affiliation – the affiliation.fuz file uses the loosest criteria for a match. As expected, it is optimised to match long place names. An ignore list is used to skip common terms and abbreviations, then it searches for terms that share most of their remaining words. For example
