Efficient Query Processing On Road Traffic Network

Road traffic data analysis has become a major focus of data mining research due to the key role it plays in optimizing road network performance. A major challenge for query processing on road traffic data is the time and memory required due to the large volume of traffic measurements available. We propose a method that provides high compression ratio while achieving high query accuracy for compressed network traffic data. We use temporal graphs for representing traffic data as they can preserve both the topological and temporal relations in the network. Unlike existing methods which mainly focus on unweighted graphs, we aim at analysing connected weighted graphs. The main idea is to exploit the redundancies between consecutive snapshots in a temporal graph to approximate the queries efficiently. Our method is demonstrated on real life road traffic network data from Melbourne. Results are compared to the output of another recent approach, and our results show that our proposed algorithm can achieve greater time efficiency with similar accuracy.