Research On Location Problem Of City Chain Stores Based On Big Data

Location,is the choice and determination of the shops,in the increasingly competitive business market today,the location selection can be regarded as the most important problem in the business process of merchants.Because as a businessman,he can change the type of goods,the price,the type of service and promotion methods,but the shop' location once be established,it can not be changed in a very long time,the change will cost a lot of manpower and material resource,so if you can chose a location with developed traffic,large people counting,then the management efficiency will be higher,the corresponding economic benefits of the shop also will be greater.Especially for the chain store business,they consider not only the location of a shop,but also the compete with each other and if the number of stores in a same region is reasonable,so how to develop a more scientific and more accurate according to the city chain store location model is the problem of this paper want to explore.This paper makes the further research and optimization of the existing location model with the combination of big data and machine learning algorithms,different classification algorithm based on machine learning are introduced,combined with the theory of retail location to summarize the location factors have great influence on the results,especially for the influence factors of the chain store summing up,and these factors were decomposed,combined with the existing data set for feature extraction,and then use SFS algorithm for feature selection,the final purpose is to improve the recommendation accuracy.Then we take Shanghai “Hua Shi” large pharmacy branch site as the research target,we have the Shanghai city traffic data,each agency distribution related data sets were combined with machine learning in GBRT,SVM and RFC experiments,to determine the most suitable location problem in the algorithm,and the use of different features of many experiments comprehensive performance,various evaluation indexes of each algorithm were judged,we show the effectiveness of the model optimization method.This paper extends the research on the location of chain stores,which has certain research and practical value.