Using latent semantic analysis for automated keyword extraction from large document corpora

In this study, we describe a keyword extraction technique that uses latent semantic analysis (LSA) to identify semantically important single topic words or keywords. We compare our method against two other automated keyword extractors, Tf-idf (term frequency-inverse document frequency) and Metamap, using human-annotated keywords as a reference. Our results suggest that the LSA-based keyword extraction method performs comparably to the other techniques. Therefore, in an incremental update setting, the LSA-based keyword extraction method can be preferably used to extract keywords from text descriptions from big data when compared to existing keyword extraction methods.