Abstract: Question Generation (QG) and Question Answering (QA) are among the many challenges in natural language generation and natural language understanding. An automated QG system focuses on generation of expressive and factoid questions which assist in meetings, customer helpline, specific domain services, and Educational Institutes etc. In this paper, the proposed system addresses the generation of factoid or wh-questions from sentences in a corpus consisting of factual, descriptive and unbiased details. We discuss our heuristic algorithm for sentence simplification or pre-processing and the knowledge base extracted from previous step is stored in a structured format which assists us in further processing. We further discuss the sentence semantic relation which enables us to construct questions following certain recognizable patterns among different sentence entities, following by the evaluation of question generated.
