Automatic Document Metadata Extraction using Support Vector Machine

Thoroughness of coverage of document

lEvaluation number of terms identified relative to the size of the text.

¡(Unique NPs#)/ (Words# in a document) is around .27