We found a match
Your institution may have access to this item. Find your institution then sign in to continue.
- Title
Systematic Analysis of the Literature Addressing the Use of Machine Learning Techniques in Transportation—A Methodology and Its Application.
- Authors
Gal-Tzur, Ayelet; Albagli-Kim, Sivan
- Abstract
Advances in the field of machine learning (ML) have been reflected in the intensity of research studies exploiting these techniques for a better understanding of existing phenomena, and for predicting future ones, as a mean for promoting a more efficient and sustainable transportation system. The present study aims to understand the trends of utilizing diverse ML approaches to tackle issues within sub-domains of transportation and to identify underutilized potentials among them. This paper presents a methodology for the bi-dimensional classification of a large corpus of scientific articles. The articles are classified into six transport-related sub-domains, based on the definition of the Israeli Smart Transport Research Center, whose aim is a transportation system with zero externalities, and the ML techniques used in each of them is identified. A fuzzy KNN model is implemented for the multi-classification of articles into the transportation sub-domains and an ontology-based reasoning for identifying the share of each applied ML approach is employed. The application of these methodologies to a corpus of 1718 articles revealed, among other findings, an increasing share of artificial neural networks and deep learning techniques from 2018 until 2022, particularly in the traffic management sub-domain. A significant contribution of the development of these automatic methodologies is the ability to reuse them for ongoing exploration of trends regarding the use of ML techniques for transportation sub-domains.
- Subjects
ARTIFICIAL neural networks; MACHINE learning; DEEP learning; SUSTAINABLE transportation; TERMINALS (Transportation)
- Publication
Sustainability (2071-1050), 2024, Vol 16, Issue 1, p207
- ISSN
2071-1050
- Publication type
Article
- DOI
10.3390/su16010207