A couple of decades ago research on Question Answering System (QAS) was started. Presently, there are many QASs available in the world. Still there is a need for development of QAS. The web is increasingly becoming an idle source of answer to different questions because of the tremendous amount of information that is now available online. The paper provides a complete overview of QAS. It presents the Question Answering (QA) task from an information retrieval perception and emphasizes the significance of retrieval models. The study suggests the architecture of QAS, gradually increasing the complexity of representation level of question and information object.

The main task of Question Answering (QA) is providing a short answer to natural language query supported by a document in a core text collection. Many Question Answering Systems (QASs) approach the problem from information extraction angle. There is a need to solve this problem to retrieve the information correctly. For these problems, the QAS is introduced to overcome and bring effective answer for userrequested queries. The goal of QAS is to identify and present the user an actual answer to a question formulated in natural language (Kepei and Jieyu, 2010). In earlier days, several technologies were used in QAS. In this context, QAS aims to provide relevant information to end-users as correct answer to a subjective question through a search in both unstructured and structured data collections. There were several studies made on QA in the earlier period. Simmons (1965) reviewed the very first approaches of QA in English. Hirschman and Gaizauskas (2001) discussed the background, motivation and general approaches to open domain QA highly promoted by the Text Retrieval Conference (TREC). An update on the approaches used in open domain QA was presented in Mihai et al. (2003).

Information Technology Journal, Question Answering (QA), Natural language processing.