Abstract: City’s functional regions refer to the geographical regions with relatively fixed functions(such as industry,commerce,housing,education,etc.) in the development of city.The position structure of these functional regionsaffect people’s travel patterns,and these travel patterns also objectively reflect the real function of regions.This paper focused on the travel patterns of urban residents by using the taxicabs’ trajectory data,and obtained functional regions accor-ding to these travel models.The main contributions of this paper are as follows.Firstly,this paper constructed the region pattern graph by using the taxicabs’ trajectories and the road network structures,and then connected different geographical regions via the graph structure created by region graph pattern construcing algorithm.Secondly,this paper proposed the framework and basic implementation idea of bottom-up functional region discovering algorithm,including mining the frequent travel pattern subgraphs and discovering frequent travel pattern from these subgraphs.Thirdly,this paper proposed a functional region cluster algorithm to cluster the obtained travel pattern graph set,thus discovering the functional regions according to the clustering results.The experimental results show that this method is effective and achieves higher purity of the function compared with traditional methods,and the entropy is decreased by 10%.

Key words: City big data, City functional region, Data mining, Travel pattern subgraph