Master of Science in Engineering (MSE)
Civil and Environmental Engineering
First Committee Member
Second Committee Member
Third Committee Member
Fourth Committee Member
Fifth Committee Member
Number of Pages
Performance measure is a process of evaluating and quantifying a system. Performance measure provides us with information about how good a system is working and how well the predefined goals are met. In order to analyze the performance of a transportation system, the traffic data such as speed, volume, occupancy and travel time of the system need to be collected. These data will generate valuable historical database that can be used to develop models to improve the quality of service of transportation system. The performance measures in transportation studies can be categorized to following main groups: Congestion, Mobility, Accessibility, Reliability, Safety and Environmental. Traffic congestion is one the important issues in any transportation system. Growing congestion in urban transportation network has enforced significant economic burdens to our current society. It causes waste of time, money, fuel and energy for the commuters and consequently impacting daily life of people in the society. Based on 2011 Congested Corridors Report presented by Texas A& M Transportation Institute, traffic congestion incurred $121 billion cost for drivers. Based on this report, 5.5 billion additional hours are wasted waiting in traffic in 2011. It means $818 additional fuel and time cost for each commuter. Being aware of the status of congestion in future can help, decision makers, intelligent systems and apps improve their accuracy and help commuters in their travel routing. To achieve these goals accurate traffic status classification techniques is required. Achieving higher accuracy is still one of the influential driving factor for research in this area. The objective of this thesis is to utilize data mining techniques to classify traffic status to congested or non-congested for some point of time in future based on historical traffic parameters (Vehicle Count, Occupancy, Speed). Moreover, to compare the performance of different data mining techniques on this problem. This dissertation examined several classification techniques including J48 Decision Tree, Artificial Neural Network, Support Vector machine, PART and K-Nearest Neighborhood to classify future traffic status to Congested or Non-congested. The one minute traffic data from I-15 Northbound from I-215 up to Desert Inn, Las Vegas, NV were used to run these experiments. Based on the comparison of these algorithms, the J48 algorithm has the best performance.
Congestion prediction; Data mining; Traffic – Computer simulation; Traffic congestion; Traffic congestion – Prevention; Traffic engineering
Civil and Environmental Engineering | Transportation
University of Nevada, Las Vegas
Mirakhorli, Abbas, "A Comparative Study: Utilizing Data Mining Techniques to Classify Traffic Congestion Status" (2014). UNLV Theses, Dissertations, Professional Papers, and Capstones. 2197.
IN COPYRIGHT. For more information about this rights statement, please visit http://rightsstatements.org/vocab/InC/1.0/