Transit electrification state of the art: A machine-learning based text mining approach

Despite the plethora of studies on electric buses, there remains a lack of knowledge synthetization that brings together previous findings and unveils latent patterns of the literature. That said, this study aims to develop a knowledge synthesis model for e-bus literature by identifying key latent research topics and patterns as well as characterizing both emerging and declining topics. The study utilized a dataset of 340 abstracts of e-bus literature extracted from the Web of Science. Furthermore, the study employed a topic modelling approach, namely the Latent Dirichlet Allocation, to analyze the literature from 2000 to 2021. The analysis identified 30 latent research topics in the literature, covering the main research areas of e-bus development, adoption, and operation. In addition, the results highlighted the most emerging research topics and areas for future research. Most importantly, the lack of intersectionality between key research topics is identified to inform future research direction.