Automatically Identifying Themes and Trends in Software Engineering Research

• Main Authors: Kenneth Cosh, Sakgasit Ramingwong, Narissara Eiamkanitchat, Lachana Ramingwong
• Format: Conference Proceeding
• Published: 2018
• Subjects: Computer Science; Mathematics
• Online Access: https://www.scopus.com/inward/record.uri?partnerID=HzOxMe3b&scp=85052334429&origin=inward; http://cmuir.cmu.ac.th/jspui/handle/6653943832/58483
• Institution: Chiang Mai University

© 2018 IEEE. Understanding the ways that research topics are evolving in a research domain is important when considering research proposals. Bibliometric analysis provides a variety of tools for exploring publication data, but often involves manual effort. This paper presents an automatic method for extracting and examining key research themes by using natural language processing to parse a large collection of papers. The method was applied to over 8,000 papers published in the software engineering field over the past 20 years. Key research themes were identified and visualized, so that trends could be highlighted. Some research fields that are in decline are identified, along with newly popular research topics such as fuzzy set membership, cloud computing, feature selection and agile development teams.