TY - JOUR
T1 - Comprehensive Survey
T2 - Approaches to Emerging Technologies Detection within Scientific Publications
AU - Yelenov, Amir
AU - Pak, Alexandr A.
AU - Ziyaden, Atabay A.
AU - Akhmetov, Iskander
AU - Gelbukh, Alexander
AU - Gelbukh, Irina
N1 - Publisher Copyright:
© 2022 Instituto Politecnico Nacional. All rights reserved.
PY - 2022
Y1 - 2022
N2 - The identification of breakthrough topics and emerging technologies has been of interest to the governments of many countries and the scientific community since the last century. This study presents the status and trend of the research field through a comprehensive review of relevant publications, a new look at the problem of defining the term "emergent technologies," defining boundaries between similar terms; and a modern baseline method on the citation prediction subtask for the discovery of emergent technologies. The outcomes of this technique have demonstrated the significance of features that characterize the preceding 1-year, 2-year, and 3-year citation counts, as well as their impact on the quality of neural network and random forest models. Our hypothesis, however, that author-specific measures may enhance prediction results was not supported. We ascribe this difficulty to the dimensionality curse. The authors examined methodological elements of research and technological development; consequently, it is important to note that, from a technical viewpoint, theoretical research is far from complete due to the vast variety of projects, outstanding challenges, research questions, and market assumptions. Finding more input characteristics to improve the quality of predictions and switching from classification to regression may also improve the precision of the suggested baseline model.
AB - The identification of breakthrough topics and emerging technologies has been of interest to the governments of many countries and the scientific community since the last century. This study presents the status and trend of the research field through a comprehensive review of relevant publications, a new look at the problem of defining the term "emergent technologies," defining boundaries between similar terms; and a modern baseline method on the citation prediction subtask for the discovery of emergent technologies. The outcomes of this technique have demonstrated the significance of features that characterize the preceding 1-year, 2-year, and 3-year citation counts, as well as their impact on the quality of neural network and random forest models. Our hypothesis, however, that author-specific measures may enhance prediction results was not supported. We ascribe this difficulty to the dimensionality curse. The authors examined methodological elements of research and technological development; consequently, it is important to note that, from a technical viewpoint, theoretical research is far from complete due to the vast variety of projects, outstanding challenges, research questions, and market assumptions. Finding more input characteristics to improve the quality of predictions and switching from classification to regression may also improve the precision of the suggested baseline model.
KW - Citation prediction
KW - emergent technology
KW - neural networks
KW - scientometrics
UR - http://www.scopus.com/inward/record.url?scp=85146878467&partnerID=8YFLogxK
U2 - 10.13053/CyS-26-4-4424
DO - 10.13053/CyS-26-4-4424
M3 - Artículo
AN - SCOPUS:85146878467
SN - 1405-5546
VL - 26
SP - 1587
EP - 1601
JO - Computacion y Sistemas
JF - Computacion y Sistemas
IS - 4
ER -