关键词: 时间序列, 相似性, 灰色关联, 聚类分析, 最大频繁项集

Abstract: The division of traffic road section by the traffic flow similarity plays an important role in urban traffic control and management. Traffic flow has a characteristic of time series. Similarity measurement is one of the important problems in time series clustering problems. A similarity measurement method based on grey correlation is proposed for traffic flow dada clustering. The results of comparative experiments show that the proposed method has a high accuracy of clustering. As the differences of traffic flow time series in each time period can influence the accuracy of traffic flow data clustering in the whole time period, a clustering method of traffic flow data based on the division of time period is proposed in this paper. The clustering of traffic flow data for each time period is first studied, and a traffic road section is divided into several groups by maximum frequent itemsets method. The validity of this method is tested by the actual traffic flow data.

Key words: time-series, similarity, grey correlation, clustering analysis, maximum frequent itemsets