Font Size: a A A

Research On Emerging Pattern Analysis Method Based On Urban Computing

Posted on:2021-01-05Degree:MasterType:Thesis
Country:ChinaCandidate:F Z XiaoFull Text:PDF
GTID:2428330626965631Subject:Computer Science and Technology
Abstract/Summary:PDF Full Text Request
In the wave of urban informatization and intelligence,urban data is the cornerstone of urban informatization.How to deal with and analyze urban data has become a research hotspot.Therefore,this paper makes an in-depth study on how to mine the exposure mode of urban data.Because the city data has the characteristics of multi-dimensional heterogeneity and rich sources,it is impossible to directly mine the exposed patterns.Therefore,this paper gives the corresponding data feature extraction methods according to the different data types and different mining objectives of the city,which makes it possible to mine the exposed patterns of the city data.At the same time,for the data without the pre-determined feature extraction methods,it puts forward the use of the original city data the algorithm of mining suspected exposure patterns from initial data distribution can make full use of urban data.The main work is as follows:1.Design corresponding data feature extraction methods for different types of urban data and mining objectivesIn this paper,for several common urban data,such as POI data,GPS data,public transport data,and so on,different feature extraction methods are given in combination with different mining targets.For example,the mining target of POI data in cities may be the exposure mode of different types of POI,or the significant difference attribute of POI between different functional areas.At the same time,in order to mine POI exposure patterns in different regions,a feature extraction method based on hierarchical tree to cluster POI names directly is designed.Through real city data experiments,effective exposure patterns are mined.2.Mining suspected exposure patterns for data without pre feature extractionIt is impossible to design corresponding feature extraction methods for all data and different mining targets in advance,so these data cannot be exposed to reveal patterns.In order to solve this problem,in this paper,based on lattice algebra theory,for the state of urban original data distribution,look for in the lattice Labeling the dominant nodes of the data,an algorithm s-EP for mining suspected revealing patterns in the area outside the dominant data points is proposed,and the effectiveness of the algorithm is analyzed.
Keywords/Search Tags:Emerging pattern, Urban computing, Multi-source heterogeneous data, data cube, Suspected emerging pattern
PDF Full Text Request
Related items