Domain-independent automatic keyphrase indexing with small training sets

摘要: Keyphrases are widely used in both physical and digital libraries as a brief, but precise, summary of documents. They help organize material based on content, provide thematic access, represent search results, assist with navigation. Manual assignment is expensive because trained human indexers must reach an understanding the document select appropriate descriptors according to defined cataloging rules. We propose new method that enhances automatic keyphrase extraction by using semantic information about terms phrases gleaned from domain-specific thesaurus. The key advantage approach it performs well very little training data. evaluate large set manually indexed documents domain agriculture, compare its consistency group six professional indexers, explore performance smaller collections other domains French Spanish © 2008 Wiley Periodicals, Inc.