Automatic Question Answering Systems: A Comprehensive Review

Abstract

Users can use search engines such as Google and Yahoo! to search for documents on the World Wide Web. It takes time, but the user of a search engine must go through each document to find an answer that is relevant to the question. The Query Answering (QA) method reduces the amount of time spent searching for the exact answer to a question. The study of question-answering systems is an important aspect of the field of information retrieval. The year 1960 saw the start of research into question-answering systems, and since then, a plethora of different question-answering systems have been developed. The Question Answering system combines research from several fields, including Natural Language Processing, Artificial Intelligence, Information Retrieval, and Information Extraction. The goal of a question answering system is to provide a precise response in natural language to the user\'s question. The availability of various resources for responses is used to differentiate between different types of question answering systems. In comparison to the open domain question answering system, the closed domain question answering system provides more precise and accurate responses.

Introduction

Automatic Question Answering (QA) systems are designed to provide direct and accurate answers to questions asked in natural language, unlike traditional search engines that return a list of web pages requiring users to search for relevant information manually. QA systems integrate Natural Language Processing (NLP), Information Retrieval (IR), Information Extraction (IE), and Artificial Intelligence (AI) to understand user queries and generate precise answers, making information retrieval faster and more user-friendly.

The literature review highlights various techniques used in QA systems, including Jaccard similarity, BERT, Deep Learning, Support Vector Machines (SVM), Random Forest, TF-IDF, ontology parsing, Named Entity Recognition (NER), clustering, and template-based methods. Although these techniques improve answer accuracy, researchers have identified several limitations such as poor handling of synonyms, lack of multilingual support, manual template creation, limited datasets, difficulty in answering unanswerable questions, and the need for improved intent identification and speech recognition.

QA systems are broadly classified into Open-Domain and Closed-Domain systems. Open-domain QA systems answer questions on any topic using large knowledge sources such as the internet and Wikipedia, but the quality of responses may vary. Closed-domain QA systems focus on a specific subject area such as healthcare, patents, or geography. They provide more accurate answers because they use specialized knowledge bases, ontologies, and structured datasets.

Questions handled by QA systems are categorized into several types:

• Factoid Questions: Require short factual answers such as names, dates, places, or organizations.
• List Questions: Return multiple relevant items or names.
• Confirmation Questions: Require yes/no answers and depend on reasoning and world knowledge.
• Causal Questions: Seek explanations or reasons and require descriptive responses.
• Hypothetical Questions: Involve speculative scenarios with multiple possible answers.
• Complex Questions: Require information synthesis from multiple sources and detailed explanations.

A typical QA system follows three main stages:

1. Question Analysis: Identifies the type of question, extracts keywords, and determines the expected answer type.
2. Passage Retrieval: Searches the knowledge base using techniques such as TF-IDF to retrieve the most relevant documents or passages.
3. Answer Extraction: Selects and ranks the most relevant answer using similarity measures and ranking algorithms.

The paper also discusses three major approaches for developing QA systems:

• Linguistic Approach: Uses NLP techniques such as tokenization, parsing, part-of-speech tagging, semantic analysis, ontologies, and knowledge bases to understand language and retrieve accurate answers. While highly accurate within specific domains, it requires extensive domain-specific knowledge.
• Statistical Approach: Employs machine learning algorithms such as SVM, Bayesian classifiers, Maximum Entropy models, and statistical text analysis. These methods can process large datasets and adapt to new domains but may overlook deeper linguistic relationships.
• Pattern Matching Approach: Matches user questions with predefined text patterns or templates. This approach is simple, efficient, and suitable for small or medium-sized systems. It includes surface pattern-based methods, which use text patterns to identify answers, and template-based methods, which use predefined question and answer templates with dynamic parameter filling.

Conclusion

In this survey, an overview of the structure of a question answering system is discussed. This article describes several different kinds of Question Answering systems. A closed-domain Q&A system restricts questions to those within a specific topic area and ensures the quality of the answers provided. Whereas, in the public domain, question and answer systems are available, which end up providing subpar results as they are not trained for any particular area. Here, we investigate how a Question Answering system deals with inquiries. To enhance the natural language processing, it is expected to work upon a few points. Initially, the synonyms checking logic, grammatical structure of the sentence, and importance decision of each sentence part should be improved. The system must be tested using various test procedures and datasets for getting the correct output. Another important task is dealing with questions that have no answers within the system. Adding features that allow the system to understand and respond in multiple languages would also be helpful. We should also look into making the system work with speech recognition, which would make it more accessible to people with disabilities. The limitation of the system should also be kept in mind, like the manual preparation of writing templates is laborious and the performance depends on the goodness of the search engine. Also, we should look towards making the system handle more types of questions and compare the performance with other tools like BERT. Thus, by working in this area, natural language processing systems can be made to work better and be more useful to people. We further came to know that researchers more frequently use a closed-domain question-answering system as compared to an open-domain question-answering system.

References

[1] Ali Mohamed Nabil Allam, Mohamed Hassan Haggag, “The Question Answering Systems: A Survey.”, International Journal of Research and Reviews in Information Sciences (IJRRIS) Vol. 2, No. 3, September 2012. [2] Bidyut Das, Mukta Majumder, SantanuPhadikar, Arif Ahmed Sekh,“Automatic question generation and answer assessment: a survey”, Research and Practice in Technology Enhanced Learning (2021) [3] Maria Vijoy and Sangeetha Jamal., “Question Answering System with Repeated Question Identification”, International Journal of Advanced Research (2016), Volume 4, Issue 7, 1835-1839 [4] Junichi Fukumoto ,” Question Answering System For Non-Factoid Type Questions And Automatic Evaluation Based On BE Method”, Proceedings Of NTCIR-6 Workshop Meeting, May 15-18, 2007, Tokyo, Japan. [5] Tilani Gunawardena, Nishara Pathirana, MedhaviLokuhetti, Roshan Ragel, and Sampath Deegalla, “Performance Evaluation Techniques for an Automatic Question Answering System”, International Journal of Machine Learning and Computing, Vol. 5, No. 4, August 2015 [6] Shervin Minaee, Zhu Liu, “Automatic Question-Answering Using A Deep Similarity Neural Network”, Computation and Language,5 Aug 2017. [7] ChaninPithyaachariyakul, Anagha Kulkarni , “AutomatedQuestionAnsweringSystem forCommunity-BasedQuestions”,The Thirty-Second AAAI Conference on Artificial Intelligence (AAAI-18),2018. [8] Wenxiu Xie , Ruoyao Ding , Jun Yan ,Yingying Qu, “A Mobile-Based Question-Answering and Early Warning System for Assisting Diabetes Management”,Hindawi Wireless Communications and Mobile Computing Volume 2018. [9] VishwajeetKumar,SivaanandhMuneeswaran, Ganesh Ramakrishnan, andYuan-FangLi , “ParaQG: A System for Generating Questions and Answers from Paragraphs”,Proceedings of the 2019 EMNLP and the 9th IJCNLP (System Demonstrations), pages 175–180 Hong Kong, China, November 3 – 7, 2019. [10] David Dominguez-Sal, Mihai Surdeanu,“A Machine Learning Approach for Factoid Question Answering”,Procesamiento del Lenguaje Natural, núm. 37, 2006. [11] Shivani G. Aithal, Abishek B. Rao, Sanjay Singh, “Automatic question-answer pairs generation and question similarity mechanism in question answering system”,Applied Intelligence (2021) 51:8484–8497 [12] Jafar A. Alzubi, Rachna Jain, Anubhav Singh, PriteeParwekar, Meenu Gupta , “COBERT:COVID-19Question Answering System Using BERT”,Arabian Journal for Science and Engineering,23 June 2021. [13] Tilani Gunawardena, MedhaviLokuhetti, Nishara Pathirana, Roshan Ragel and Sampath Deegalla, “An Automatic Answering System with Template Matching for Natural Language Questions”,International Conference on Information and Automation for Sustainability (ICIAfS),2010. [14] J. Gregory Caporaso, William A. Baumgartner, J, Hyunmin Kim, “Concept Recognition, Information Retrieval, and Machine Learning in Genomics Question-Answering”, Proceedings of the Fifteenth Text REtrieval Conference, TREC 2006, Gaithersburg, Maryland, November 14-17, 2006. [15] Shilpa Sharma, Prof. Manmohan Singh, “Implementation on Framework for Restricted Domain Question Answering System using Advance NLP Tools & Software in”,International Journal for Research in Applied Science & Engineering Technology (IJRASET) ISSN: 2321-9653; IC Value: 45.98; SJ Impact Factor: 6.887 Volume 6 Issue V, May 2018. [16] ShiyanOu, Constantin Orasan, Dalila Mekhaldi, Laura Hasler , “Automatic Question Pattern Generation for Ontology-based Question Answering”,Proceedings of the Twenty-First International FLAIRS Conference (2008) [17] Min-kyoung Kim, Han-joon Kim, “Design of Question Answering System with Automated Question Generation”,Fourth International Conference on Networked Computing and Advanced Information Management,978-0-7695-3322-3/08,2008 [18] Yikang Shao, Guangde Xu1, Mingxue Xu, Lanfang Dong, “An Automatic Question Answering Method for Small-Scale Corpus”, Journal of Physics: Conference Series 1621 (2020) 012113 [19] Pawan Kumar, Raj Kumar Goel , Prem Sagar Sharma, “A New Architecture of Automatic Question Answering System using Ontology”,International Journal of Computer Applications (0975 – 8887) Volume 97 – No.20, July 2014 [20] Meera Udani, AvinashShrivas, Vaibhav Shukla, Archana Rao, “Question Answering System Based on Artificial Intelligence for Restricted Domain”,International Journal of Engineering Research & Technology (IJERT),ISSN: 2278-0181,Vol. 2 Issue 12, December – 2013. [21] Shen Song, Yu-N Cheah, Enya Kong Tang, Bali Ranaivo-Malançon , “Rule Extraction for Automatic Question Answering Based on Structural Clustering”,International Journal of Computer Science and Network Security, VOL.8 No.3, March 2008. [22] Widodo Budiharto, Vincent Andreas, Alexander Agung Santoso Gunawan , “Deep learning?based question answering system for intelligent humanoid robot”,Journal of Big Data (2020) 7:77. [23] Francesca Alloatti,Luigi DiCaro ,GianpieroSportelli, “RealLife Application of a Question Answering System Using BERT Language Model”,Proceedings of the SIGDial 2019 Conference, pages 250–253 Stockholm, Sweden, 11-13 September 2019. [24] Yikang Shao, Guangde Xu, Mingxue Xu and Lanfang Dong, “An Automatic Question Answering Method for Small-Scale Corpus”,Journal of Physics: Conference Series 1621 (2020) 012113. [25] Kaur, and Gupta, V., “Effective Question Answering Techniques and their Evaluation Metrics”, International Journal of Computer Applications, 65(12):30–7,2013. [26] Pundge, A.M., Khillare, S.A., and Mahender, C.N., “Question Answering System, Approaches and Techniques: A Review”, International Journal of Computer Applications, 141(3):0975–8887, 2016. [27] Mishra,A., and Jain, S. K., “A survey on question answering systems with classification”, Journal of King Saud University. Computer and Information Sciences, 28(3), 345–61,2015. [28] Youzheng, Hori, Hisashi, “Leveraging social Q&A collections for improving complex question answering”, Computer Speech and Language, 29, 1–19,2015. [29] Chen, W., Zeng, Q., Wenyin, L., and Hao, T. , “A user reputation model for a user-interactive question answering system”, Concurrency Computat.: Pract. Exper., pp. 2091–103,2015 [30] Moldovan, D., Pasca, M., and Harabagiu, S.,“Performance issues and error analysis in an open-domain question answering system”, ACM Trans. Information System, 2007. [31] Kolomiyets, O., “A survey on question answering technology from an information retrieval perspective”, Inf. Sci. 181(24):5412–34., 2011. [32] Setio Basuki, Ayu Purwarianti, Statistical-based Approach for Indonesian Complex Factoid Question Decomposition, International Journal on Electrical Engineering and Informatics ,8, 2, June 2016,356-373 [33] Iftene, A, “Textual Entailment. Ph.D Thesis”, Al. I. Cuza University, Faculty of Computer Science, Iasi, Romania, 2009. [34] Hoffner, K., Walter, S., Marx, E., Usbeck, R., Lehmann, J., and Ngomo, A. N, “Survey on Challenges of Question Answering in the Semantic Web”, Proceedings Semantic Web, pp. 1–26. 2016. [35] Ravichandran, D., and Hovy, E.H., “Learning surface text patterns for a question answering system”, In Proceedings of ACL, 2002. [36] Unmeshsasikumar, Sindhu L, “A survey of Natural Language question answering system”, international journal of computer applications (0975-8887), volume 108 -No 15. December 2014 [37] Sanjay K. Dwivedi and vaishalisingh, “Research and reviews in question answering system”, International Conference on Computational Intelligence: Modeling Techniques and Applications CIMTA) 2013 Procedia Technology 10,2013. [38] Zhang, D., and Lee, W. S., “Web based pattern mining and matching approach to question answering”, In Proceedings of the 11th Text Retrieval Conference, 2002. [39] Sneiders E, “Automated question answering using question templates that cover the conceptual model of the database”, In Natural Language Processing and Information Systems, Springer Berlin Heidelberg, 2002.

Copyright

Copyright © 2026 Sanah Nashir Sayyed, Bharat Shelke, C. Namrata Mahender. This is an open access article distributed under the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.