TY - GEN
T1 - Towards mobility-based clustering
AU - Liu, Siyuan
AU - Liu, Yunhuai
AU - Ni, Lionel M.
AU - Fan, Jianping
AU - Li, Minglu
PY - 2010
Y1 - 2010
N2 - Identifying hot spots of moving vehicles in an urban area is essential to many smart city applications. The practical research on hot spots in smart city presents many unique features, such as highly mobile environments, supremely limited size of sample objects, and the non-uniform, biased samples. All these features have raised new challenges that make the traditional density-based clustering algorithms fail to capture the real clustering property of objects, making the results less meaningful. In this paper we propose a novel, non-density-based approach called mobility-based clustering. The key idea is that sample objects are employed as "sensors" to perceive the vehicle crowdedness in nearby areas using their instant mobility, rather than the "object representatives". As such the mobility of samples is naturally incorporated. Several key factors beyond the vehicle crowdedness have been identified and techniques to compensate these effects are proposed. We evaluate the performance of mobility-based clustering based on real traffic situations. Experimental results show that using 0.3 % of vehicles as the samples, mobility-based clustering can accurately identify hot spots which can hardly be obtained by the latest representative algorithm UMicro.
AB - Identifying hot spots of moving vehicles in an urban area is essential to many smart city applications. The practical research on hot spots in smart city presents many unique features, such as highly mobile environments, supremely limited size of sample objects, and the non-uniform, biased samples. All these features have raised new challenges that make the traditional density-based clustering algorithms fail to capture the real clustering property of objects, making the results less meaningful. In this paper we propose a novel, non-density-based approach called mobility-based clustering. The key idea is that sample objects are employed as "sensors" to perceive the vehicle crowdedness in nearby areas using their instant mobility, rather than the "object representatives". As such the mobility of samples is naturally incorporated. Several key factors beyond the vehicle crowdedness have been identified and techniques to compensate these effects are proposed. We evaluate the performance of mobility-based clustering based on real traffic situations. Experimental results show that using 0.3 % of vehicles as the samples, mobility-based clustering can accurately identify hot spots which can hardly be obtained by the latest representative algorithm UMicro.
UR - http://www.scopus.com/inward/record.url?scp=77956212962&partnerID=8YFLogxK
UR - http://www.scopus.com/inward/citedby.url?scp=77956212962&partnerID=8YFLogxK
U2 - 10.1145/1835804.1835920
DO - 10.1145/1835804.1835920
M3 - Conference contribution
AN - SCOPUS:77956212962
SN - 9781450300551
T3 - Proceedings of the ACM SIGKDD International Conference on Knowledge Discovery and Data Mining
SP - 919
EP - 927
BT - KDD'10 - Proceedings of the 16th ACM SIGKDD International Conference on Knowledge Discovery and Data
T2 - 16th ACM SIGKDD International Conference on Knowledge Discovery and Data Mining, KDD-2010
Y2 - 25 July 2010 through 28 July 2010
ER -