QUESTION ANSWERING SYSTEMS

The question answering systems are a way of information retrieval that responds to the question asked in natural language. They consist of three main components : question classification, information retrieval and response extraction. The implementation of large-scale joint assessments such as TREC conference, has created a powerful interest community and accelerated the progress in the study of question answering systems. This paper presents the problem domains in answering questions. It reviews the historical development of the question answering systems, from the simplest highly specialized ones to the present more sophisticated and higer quality systems that are able to give us short and concise answers to questions belonging to a variety of domains. The systems are analyzed according to the purpose. The paper provides an overview of previous research on question answering systems. It defines the general architecture of question answering systems and analyzes current approaches to every phase of architecture. The question answering systems are still under development and experimentation. In conclusion, the selection of the most accurate and the most concise response that will satisfy the user, still remains the largest problem in question answering systems.