The Research And Application On The Model Of Urban Functional Features Based On Mobile Data

With the continuous expansion of urban scale and the change of the industrial structure,the internal function of the city tends to polarise into different functional areas,such as residential areas,commercial areas and business areas.How to effectively identify these functional areas and extract their distribution features is an important prerequisite for rational planning of urban space.Currently,the frequently-used identification methods include questionnaire survey,remote sensing and GPS portable detection.But these methods are often accompanied with problems like long running cycle,narrow coverage,high cost and so on.Compare with that,the abundant mobile data have many excellent properties such as low cost,broad coverage and good real-time performance.Considering all of this,the method of analyzing and recognizing urban functional features based on mobile data is proposed,and the features model of urban functional area is constructed.Urban functional areas studied in this thesis are residential areas,commercial areas and business areas.The data collected from the operator users are used to carry out urban refinement research and to recognize the features of urban functional areas.At the same time,a method to construct the model of urban functional features is proposed.Firstly,the research area of urban function is selected as the target area and observation area;Secondly,it is proposed to recognition users‘ stay point in different periods in the target area based on ―distance-time‖ double-layer clustering method,calculating the stay time and the resident number.And then according to the applicability of the mobile data to extract travel features to statistic the information of travel population,times and time distribution in the target area.Then,an adjustable bandwidth kernel density method is proposed to calculate the population density based on the uneven distribution of base station.Finally,the users‘ behavior of working days and weekends in different functional areas are studied from two aspects of time and space according to the features of statistical analysis,and select appropriate function features.The model of urban functional festures is constructed by fuzzy C-means algorithm.The data set used in this subject is the mobile users‘ data during one week provided by a certain operator in Chongqing.It can realize the stay point recognition,travel features statistics and population density calculation in the study area of each moment.Select functional attribute value according to the real-change of features,and construct the model to recognition urban functional area.The accuracy of stay point recognition should be reach by 80% at least,and functional recognition accuracy is not less than 75%.Finally,using the constructed model and point of interest(POI)data to recognition the functional areas of observation area,and the results are compared to verify the model.Experimental result shows that the recognition of urban functional areas proposed in this thesis can meet the design requirements of the model,which has certain practical value.