Application Research On Clustering And Balance Of Municipal Performance

With the continuous expansion of city size,the continuous improvement of city management requirements,it is imminent to improve the modernization level of city management,and the smart city came into beings.However,the digital city management is base of the smart city.At present,many digital city managements of domestic cities have been completed and have been operating a lot of time,and greatly improve the level of city management,and it has accumulated a amount of data.However,How to use these data for further improvement of the management efficiency of the digital city management? it is important to establish the scientific and reasonable performance evaluation system,and build a scientific decision-making method,and it uses historical data of the digital city management.This is the background of this thesis.First,the paper introduces the status of the digital city management,analyzes the current problems in digital city management,and gives an overview of technology which paper has been used.Further more,giving a review of the clustering algorithm,clustering similarity measurement,and clustering validation,and analyzing the K-means clustering algorithm are sensitive to initial centers,so we need to give the numbers of clusters and introduce the Kruskal algorithm and the Silhouette Coefficient to improve the effect of the K-means clustering algorithm,and use the Glass Identification data set which in the UCI data set to justify the reliability and adaptability of the K-means clustering algorithm which have been changed.Once more,giving a review of the balance and outlier detection,analyzing the LOF outlier detection algorithm are not good to detect the collective outliers,so we introduce the DBSCAN algorithm to improve the effect of the LOF outlier detection algorithm,and use the Iris data set which in the UCI data set to justify the LOF outlier detection algorithm,which have been changed and can detect the point outliers,and the collective outliers from globalization and localization.Paper uses the digital city management of one of the zone of Chongqing as the application cases.First,we analyze the scene of the digital city management,and design the software architecture,the function structure and the data structure.Further more,accounting for data characteristics of the database,we use the ETL technology for data extraction,cleaning,transformation,and load the data into the data warehouse.At last,we use the OLAP technology,the K-means clustering algorithm which has been changed and the LOF outlier detection algorithm which has been changed into multi-dimensional analysis,cluster analysis and balance analysis.So the results can help municipal management workers make decision.