TY - JOUR
T1 - The evolution of research in resources, conservation & recycling revealed by Word2vec-enhanced data mining
AU - Zhu, Jun Jie
AU - Ren, Zhiyong Jason
N1 - Funding Information:
This work was supported by the Andlinger Center for Energy and the Environment at Princeton University . We are grateful for the insightful feedback provided by Prof. Ming Xu.
Publisher Copyright:
© 2023
PY - 2023/3
Y1 - 2023/3
N2 - Resources, Conservation & Recycling (RCR) publishes original research in technological, economic, institutional, policy, and system-wide aspects of resource management and sustainability. Here we developed natural language processing (NLP) and Word2vec-based techniques to reveal for the first time the underlying patterns, interactions of research topics, and topical vectors and their connections with 10 resources covered in RCR based on all 4884 articles published since its inception. The 49 most trending-up, specific topics can be arranged into nine groups: general resources and materials, waste-based resources and materials, industrial management, human behaviors, analyzing methodologies, sustainable economy, climate-relevant, other sustainability, and other problems. The Word2vec-RCR model exhibits the distribution of topic vectors based on t-SNE, and the topics can be visually grouped into 13 major regions. The newly developed Word2vec model is proven to be effective for understanding evolution and interactions between research topics and defined subjects in RCR and broader environmental domains.
AB - Resources, Conservation & Recycling (RCR) publishes original research in technological, economic, institutional, policy, and system-wide aspects of resource management and sustainability. Here we developed natural language processing (NLP) and Word2vec-based techniques to reveal for the first time the underlying patterns, interactions of research topics, and topical vectors and their connections with 10 resources covered in RCR based on all 4884 articles published since its inception. The 49 most trending-up, specific topics can be arranged into nine groups: general resources and materials, waste-based resources and materials, industrial management, human behaviors, analyzing methodologies, sustainable economy, climate-relevant, other sustainability, and other problems. The Word2vec-RCR model exhibits the distribution of topic vectors based on t-SNE, and the topics can be visually grouped into 13 major regions. The newly developed Word2vec model is proven to be effective for understanding evolution and interactions between research topics and defined subjects in RCR and broader environmental domains.
KW - Natural language processing
KW - Research interconnection
KW - Resource
KW - Sustainability
KW - Topical evolution
KW - Word embedding
UR - http://www.scopus.com/inward/record.url?scp=85146537443&partnerID=8YFLogxK
UR - http://www.scopus.com/inward/citedby.url?scp=85146537443&partnerID=8YFLogxK
U2 - 10.1016/j.resconrec.2023.106876
DO - 10.1016/j.resconrec.2023.106876
M3 - Article
AN - SCOPUS:85146537443
SN - 0921-3449
VL - 190
JO - Resources, Conservation and Recycling
JF - Resources, Conservation and Recycling
M1 - 106876
ER -