Undercovering research trends: Network analysis of keywords in scholarly articles

In this study we present a network approach to uncovering trends in an area of research by analyzing keywords appearing in scholarly articles. We represent keywords as nodes. We link a pair of keywords if they appear in the same article. Each link is assigned a weight, representing the number of co-occurrences of the pair in different articles. We perform a statistical and visual analysis of the network's structural and temporal characteristics. These characteristics or patterns provide a broad understanding of how keywords are organized and research areas evolved over time. Our findings show that keywords organize themselves into three categories: topical keywords, complimentary keywords and, diverse keywords. Through comparative analysis of the networks built from articles published in two successive time windows, we are able to detect some interesting keyword patterns and emerging areas. Results from this analysis can be used for identifying emerging research areas or updating academic programs and course curricula. To demonstrate the approach we use keywords from the European Journal of Operational Research.