主催: Eastern Asia Society for Transportation Studies
p. 320
This paper proposes a methodological framework for developing a real-time intelligent decision support system for urban traffic congestion emergency response. To support real-time decision making, the system is required to have the capability to provide an efficient organization of input data and inferred knowledge from all kinds of data sources so as to guarantee the adaptability to changes and the reusability for different congestion situations. For this requirement, we attempt to combine data warehouse with data mining methods in building the framework. Data warehouse provides a well-organized information source for traffic congestion dispersion by means of data collection, analysis, disposal and storage. Furthermore, data mining techniques deal with the set of mixed numeric and non-numeric congestion data and information. In this way, knowledge that represents cause-and-effect dependencies among the congestion attributes are extracted.