| Question asked
|
The authors claim that "similarity feature … The authors claim that "similarity features can be chosen to be language-independent." Do they test this claim? The example fields given (ISBN and dates) seem like they would provide very little information (dates), or would only be applicable in specific cases where the medium in question has unique identifiers (ISBN) and is used with both thesauri. Are there enough useful language-independent features in general to make this a useful approach? in general to make this a useful approach?
|
