Please use this identifier to cite or link to this item: https://www.um.edu.mt/library/oar/handle/123456789/53079
Full metadata record
DC FieldValueLanguage
dc.date.accessioned2020-03-25T12:59:15Z-
dc.date.available2020-03-25T12:59:15Z-
dc.date.issued2019-
dc.identifier.citationSaliba, M. (2019). Pin pointing pain points : vehicular traffic flow intensity detection and prediction through mobile data usage (Master's dissertation).en_GB
dc.identifier.urihttps://www.um.edu.mt/library/oar/handle/123456789/53079-
dc.descriptionM.SC.ARTIFICIAL INTELLIGENCEen_GB
dc.description.abstractIn this research we propose a novel approach, consisting of an ensemble of datamining and machine learning techniques, to prove that it is possible to extract and predict vehicular traffic patterns from mobile usage data. An anonymized mobile phone usage dataset from a telecommunications provider in Malta was used to generate an origin-destination (OD) matrix that defines the top two activity hubs through clustering. The OD matrix was used to infer user trips over fastest routes between these top two locations across time. We then applied spatial binning techniques to deduce the aggregate distribution of traffic load on the traffic network. A predictive model based on an artificial neural network (ANN) was trained with the whole network traffic flow load in a time series to predict traffic level for specific nodes. Daily trip distribution showed to have a very strong correlation (r = 0.94, p < 1.1e − 11) with those reported in 2010 in the National Household Travel Survey (NHTS). Similarly a significantly strong linear relationship (r = 0.69, p < 0.001) was found when comparing mean hourly route trip delays with mean trip delay estimation recorded with a Google API. To evaluate the traffic flow count method, we compared our results with manual counts retrieved from a 2016 study by Nigel Pace. Strong instances of correlation (r = 0.75,p < 0.05) were observed for low congested traffic points. These contrasted with the weak negative correlation (r = 0.45,p < 0.05) for traffic flow in locations where traffic congestion occurs frequently. Traffic flow prediction through an ANN proved to be efficient with F1-Scores ranging from 0.58 to 0.9 for different road segments in experimentation. The proposed solution needs improvement by adding a dynamic traffic assignment to the whole algorithm. This would give more accurate results, especially for traffic flow points that tend to be congested, by capturing user route selection changes and get more precise localization of delay causes.en_GB
dc.language.isoenen_GB
dc.rightsinfo:eu-repo/semantics/openAccessen_GB
dc.subjectTraffic flow -- Maltaen_GB
dc.subjectData mining -- Maltaen_GB
dc.subjectMachine learningen_GB
dc.subjectNeural networks (Computer science) -- Maltaen_GB
dc.titlePin pointing pain points : vehicular traffic flow intensity detection and prediction through mobile data usageen_GB
dc.typemasterThesisen_GB
dc.rights.holderThe copyright of this work belongs to the author(s)/publisher. The rights of this work are as defined by the appropriate Copyright Legislation or as modified by any successive legislation. Users may access this work and can make use of the information contained in accordance with the Copyright Legislation provided that the author must be properly acknowledged. Further distribution or reproduction in any format is prohibited without the prior permission of the copyright holder.en_GB
dc.publisher.institutionUniversity of Maltaen_GB
dc.publisher.departmentFaculty of Information and Communication Technology. Department of Artificial Intelligenceen_GB
dc.description.reviewedN/Aen_GB
dc.contributor.creatorSaliba, Maurice-
Appears in Collections:Dissertations - FacICT - 2019
Dissertations - FacICTAI - 2019

Files in This Item:
File Description SizeFormat 
19MAIPT012.pdf4.27 MBAdobe PDFView/Open


Items in OAR@UM are protected by copyright, with all rights reserved, unless otherwise indicated.