Clustering of Urban Road Paths; Identifying the Optimal Set of Linear and Nonlinear Clustering Features
Ημερομηνία
2021Γλώσσα
en
Λέξη-κλειδί
Επιτομή
Urban traffic is undoubtedly a dynamic phenomenon presenting variations over both time and space, that in the majority of cases are the result of a mixture of, either well known (i.e. weather, seasonality) or not easily predictable (i.e. events, accidents) external factors. Identification of similarities in the performance of different urban road paths under different traffic states (different travel demand conditions) is the main subject of the current paper. Floating taxi travel time data (timeseries per road path) collected in the framework of Thessaloniki Smart Mobility Living Lab (initiated and operated by CERTH/HIT) consist the basic input for the hierarchical clustering that is applied. Clustering applies upon different combinations of road paths’ features (data points of travel time timeseries, descriptive statistics and mutual information of timeseries). The comparison of the clustering results based on average weekdays travel times per road path (from a six months period) with the respective results of a typical and an atypical day adds on the interpretability of underlying relations among paths under different states. The analysis reveals that resulting clusters can be a building block for the spatiotemporal understanding of urban traffic. Furthermore, it is shown that adding as clustering feature the criterion of mutual information of timeseries, therefore taking into account also non-linear dependences of the different road paths, the clustering interpretability is differentiated. © 2021, The Editor(s) (if applicable) and The Author(s), under exclusive license to Springer Nature Switzerland AG.
Collections
Related items
Showing items related by title, author, creator and subject.
-
A Scalable Short-Text Clustering Algorithm Using Apache Spark
Akritidis L., Alamaniotis M., Fevgas A., Bozanis P. (2021)Short text clustering deals with the problem of grouping together semantically similar documents with small lengths. Nowadays, huge amounts of text data is being generated by numerous applications such as microblogs, ... -
Online clustering of distributed streaming data using belief propagation techniques
Halkidi, M.; Koutsopoulos, I. (2011)Extraction of patterns out of streaming data that are generated from geographically dispersed devices is a major challenge in data mining. The sequential, distributed fashion in which data become available to the decision ... -
Distributed clustering in vehicular networks
Maglaras, L. A.; Katsaros, D. (2012)Clustering in vanets is of crucial importance in order to cope with the dynamic features of the vehicular topologies. Algorithms that give good results in Manets fail to create stable clusters since vehicular nodes are ...