Speed Transition Matrix Feature Extraction for Traffic State Estimation Using Machine Learning Algorithms (CROSBI ID 72876)
Prilog u knjizi | izvorni znanstveni rad | međunarodna recenzija
Podaci o odgovornosti
Tišljarić, Leo ; Ribić, Filip ; Majstorović, Željko ; Carić, Tonči
engleski
Speed Transition Matrix Feature Extraction for Traffic State Estimation Using Machine Learning Algorithms
Book cover The Science and Development of Transport—ZIRP 2021 pp 61–74Cite as Speed Transition Matrix Feature Extraction for Traffic State Estimation Using Machine Learning Algorithms Leo Tišljarić, Filip Ribić, Željko Majstorović & Tonči Carić Chapter First Online: 08 April 2022 Abstract Feature extraction is a crucial part of data preparation when using machine learning algorithms, especially for emerging datasets. The speed transition matrix (STM) emerged only recently as a traffic data modeling technique. In this paper, key features from STMs are extracted and proposed for the purpose of traffic state estimation. This step simplifies the learning process and the interpretability of the results obtained when estimating the traffic state using the STMs. Using the proposed features, traffic state is estimated for the most crucial road segments in the City of Zagreb, Croatia. The method is evaluated on some of the most used machine learning algorithms, with the highest accuracy value obtained with decision tree and random forest algorithms.
Speed transition matrix ; Machine learning ; Feature extraction ; Traffic state estimation
nije evidentirano
nije evidentirano
nije evidentirano
nije evidentirano
nije evidentirano
nije evidentirano
Podaci o prilogu
61-74.
objavljeno
10.1007/978-3-030-97528-9_5
Podaci o knjizi
The Science and Development of Transport—ZIRP 2021
Petrović, Marijana ; Novačko, Luka ; Božić, Dijana ; Rožić, Tomislav
Cham: Springer
2022.
978-3-030-97527-2
Povezanost rada
Informacijske i komunikacijske znanosti, Interdisciplinarne tehničke znanosti, Tehnologija prometa i transport