Research On Shopping Center Location Based On Points Of Interest

As a specific form of shopping,leisure and entertainment,the Shopping Center has been favored by people for its complete range,quality assurance,and convenience.Shopping malls are large-scale commercial real estate projects,with large buildings,large investment funds,long investment cycles,and relatively low return on capital.The shopping center address selection directly determines the operating efficiency of the shopping center.It is of great significance to use the spatial data mining technology to research the problem of shopping center location.First of all,aiming at the problem that the location of shopping centers is currently subjectively determined on the weight of site selection indicators,a random forest shopping center location model based on the Cos-POICompare algorithm is proposed.The main idea of the Cos-POICompare algorithm is the cosine similarity principle.Considering the different types of points of interest within 5km of the shopping center and the differences in location factors of shopping centers,the similarity index between shopping centers and different types of interest points in location selection is calculated.The random forest algorithm analyzes the similarity index of different types of points of interest,the GDP data of the urban area of the shopping center,the population data,and the annual sales performance of the shopping center,and realizes the prediction of the shopping center location.Second,aiming at the problem that Cos-POICompare algorithm can not take into account the distant customers,a random forest shopping center location model based on point of interest association algorithm is proposed.Considering points of interest in the city scope,point of interest association algorithm analyzes the interest degree between shopping centers and different point of interest types.The random forest algorithm analyzes the interest degree of different point of interest types,the GDP data of the city where the shopping center is located,population data,and the annual sales performance of the shopping center,and completes the prediction of the shopping center location.Finally,the Cos-POICompare algorithm and point of interest association algorithm are used to preprocess the indicators,and the selected indicators are used to predict and analyze the existing shopping centers to verify the correctness of the above two models.