Answering factoid questions via Ontologies : a natural language generation approach

Abstract

We present a systematic approach to the generation of natural language descriptions of logical facts from ontologies. We design and discuss our Natural Language Generation (NLG) architecture in terms of implementing a factoid question answer platform upon ontologies; who identify what questions can be asked upon the knowledge base, determine relevant contents from the knowledge base that best serve generating response to the questions and process those contents confirming to the popular patterns of expression, as identified from a survey, in order to generate answers in natural language (English); all of this while justifying the rationale of the approach and the possible benefits such systems can offer.