Improving Keyphrase Extraction from Biomedical Documents Using Domain Specific Feature Set

Keyphrases enable the reader to quickly determine whether the given article is suitable for the reader’s digest. Keyphrases are also important for medical document retrieval and text mining research. Sometimes, the author-assigned Keyphrases or keywords available with the articles are too limited to represent the topical content of the articles. Many medical documents also do not come with author assigned Keyphrases. So, an automatic Keyphrase extraction process is highly desirable. Domain specific features can play an important role in domain specific Keyphrase extraction. This paper presents an automatic Keyphrase extraction method that exploits a number of domain-specific features to improve the Keyphrase extraction performance in medical domain.