Analyzing scientific research topics in manufacturing field using a topic model

We can gain a thorough understanding of an academic field by surveying related scientific literature. We applied the latent Dirichlet allocation (LDA) method and developed a topic model to analyze changes in research topics over time and identify the temporal dynamics of manufacturing-related research with respect to journals and countries/areas. The data were collected from 82,248 published journal abstracts published in 43 journals between 1990 and 2016. We found that the research scope of manufacturing has become increasingly interdisciplinary; management science and materialogy have become more important. Also, manufacturing research is no longer limited to traditional sub-fields such as manufacturing processes, mechanical transmission, and machining tools but is also concerned with advanced manufacturing techniques (e.g., laser-based manufacturing, semiconductor fabrication, and planning of production processes). By quantifying the temporal trends of topics at the country/area level, we found that researchers in different regions tend to focus on different sub-fields, while adjacent countries/areas commonly share research topics. Our findings could be of great value to researchers, enterprises, journals and publishers who are interested in identifying promising or popular topics.