Publication

Abstract : Understanding natural language text by automated systems is becoming popular as the need for conversion of unstructured text to structured text increases. We report the design of a scalable Natural Language Report Management System in which information is collected from unstructured texts is done using statistical natural language processing tools like NLTK (Bird S, 2009), Stanford CoreNLP (Stanford CoreNLP, 2013) etc. The extracted information is then stored in a graph database to form the knowledge base. More information is added to this knowledge base using semantic web technology, DBpedia, and Geo-ontology. Reasoners like Pellet are used to improve the reasoning capabilities of the system. The query processing system, with the query being in natural language, will search for an absolute match in the stored knowledge base. A Natural language generation module is integrated with the system, using which the processed query result is articulated to produce answers in natural language.