Question Answering Systems for Managing Big Data

Question Answering Systems for Managing Big Data

Sparsh Mittal (Iowa State University, USA)
Copyright: © 2014 |Pages: 7
DOI: 10.4018/978-1-4666-5202-6.ch175

Chapter Preview



Zweigenbaum (2003) discuss the role and importance of QASs in biomedicine. In the literature, several techniques have been proposed for answering biomedical questions, such as answering by role identification (Niu et al., 2003) and document structure (Sang et al., 2005). In a study conducted with a test set of 100 medical questions collected from medical students, a thorough search in Google failed to obtain relevant documents within top five hits for 40% of the questions (Jacquemart & Zweigenbaum, 2003). Moreover, due to need of answering the question swiftly and the busy practice schedules, doctors spend less than two minutes on average for searching an answer to a question. Hence, search engines fail to fully answer most of the clinical questions (Ely et al., 1999). These research studies further confirm the importance of question answering systems.

Some QASs are closed-domain, which implies that they deal with a specific domain or accept only a restrict kinds of questions. Other QASs are open-domain which can answer multiple kinds of questions from different fields. An example of closed-domain QAS is biomedical QAS. Sondhi et al. (2007) discuss a biomedical QAS named Internet doctor (INDOC). Their system works by indexing the entire document set. The system processes the user-question to recognize the difference in significance of different parts of the query. The answers are ranked by measuring the relevance of the documents to the query.

Key Terms in this Chapter

Information Retrieval (IR): IR refers to the task of retrieving useful information from a data set.

Open-Domain QAS: An open-domain QAS can answer questions from different fields which include history, science and geography.

Multi-Modal QAS: A multi-modal QAS employs non-text input mode (such as audio-video) for entering the question, in addition to or in place of usual text input.

Search Engines: Search engines search the Web resources and provide documents in the response to the question need of the user.

Cross-Lingual QAS: Cross-lingual QAS offer the capability to provide answer, to a question posed in one language, in another language.

Multi-Lingual QAS: Multi-lingual QASs interact with the user in multiple languages.

Multi-Stream QAS: A multi stream QAS employs multiple streams for finding the answer of the question, where a stream is a small question answering system on its own.

Multi-Media QAS: A QAS which returns both text and multi-media information as the answer to the user question.

Multi-Perspective QAS: A Multi-perspective QAS answers opinion-seeking questions.

Natural Language Processing (NLP): NLP refers to analyzing, understanding, and generating languages that humans use naturally, without requiring modification to suit computer syntax.

Closed-Domain QAS: A closed-domain QAS answers questions on a specific field, such as a Biomedical QAS.

Question Answering System (QAS): QAS aims to answer the questions posed by humans in the natural language. Moreover, it provides “answers” and not merely documents.

Multi-Document Comparison QAS: A QAS which can generate a single answer by searching relevant data from multiple documents. This QAS can answer comparison type questions.

Complete Chapter List

Search this Book: