Discover trending domains using fusion of supervised machine learning with natural language processing

In this paper, a new technique is presented for mining key domain areas from scientific publications. A domain refers to a particular branch of scientific knowledge and hence largely defines the theme of any scientific research paper. The proposed technique stems from a fusion of knowledge derived from natural language processing and machine learning. Some words or phrases are extracted based on their meaning inferred by the application of preposition disambiguation. These key words or phrases are then classified as domain areas using supervised learning. Various experiments and their analyses yield concrete results validating the efficacy and application of our methodology. The fusion technique therefore extracts an interesting aspect of research from scientific text and hence propounds a hybrid methodology for deriving meaning from underlying text. This approach thus takes a definitive step in advancing text analytics.