Development of research trends evolution model for computer science for Malaysian publication

Abstract
Nowadays, there seem to be research trends done on studies that manipulate publications that utilise the text mining approach. However, most of these studies only investigated the gaps faced by existing research trends models, and the execution of text mining of bibliometric elements and the timeline windows representing the "trends" was not clarified. Thus, this study aimed to develop the conceptual model for research trends in Malaysian publications, specifically, to incorporate the text element of bibliometrics and the execution of timeline windows to identify research trends. In the context of research trends, the evolution or growth of some research area from one period to another is important. This included what has happened, what is currently happening, and predicting potential research trends that will happen in the near future for others to continue the research development. The element in the newly developed model was extracted from the literature review and adapted from one of the selected models. The new model consisted of three stages which is the first stage consisted of three elements - selecting document collection; the second stage was the selection of the bibliometric element; and the third stage was the execution of text mining, co-word analysis from the selected textual bibliometric element, the implementation of two timeline windows (fixed time and sliding time windows-timeline). Also, the execution of the third stage required aid from tools - CiteSpace. The newly developed model was tested, and data were downloaded from two databases, Scopus (10,052 publications) and Web of Science (WoS) (22,088 publications), for a duration between 1995 and 2019. This study identified that the research trend pattern became more active from 2002 onwards. Besides that, the research topic became fresher and more unconventional throughout the timelines. Research topics on artificial intelligence, network communication, and wireless sensor networks are the hottest topics and timeless. Besides that, knowledge management, internet banking, online shopping, and eCommerce were the alternative options for computer science researchers. Each timeline's evolution and blooming shows that researchers are investigating each topic thoroughly. In addition, some small topics do not appear in fixed timeline windows but instead emerge from sliding timeline windows, such as system development, shared banking service, virtual team collaboration, and internet policy. This study also captured the highlighted keywords that could give hints or appear as an initial idea for the next research journey. Experts' evaluation and validation were executed as the interpretation of experimental results require experts' expertise, experience, and views. A semi-structured interview was done with thirteen experts who have remarkable expertise in research and development. From the discussion, most experts agreed that the model could help others identify the research trends and potential new research topics emerging for future research journeys. The newly developed model could be beneficial to those who need hints for their next exploration and help those keen to understand how to execute the text mining within the bibliometric elements.
Description
Thesis (PhD. (Information Systems))
Keywords
Research—Data processing, Research—Management, Text data mining
Citation