Weighted DBSCAN Algorithm for Discovery of Spatial Features in Industries Management

In view of detection the key block of city industries in the research of spatial layout, the businesses in fuzhou city of jiangxi province were taked as research objects. Using check-in location data from sina weibo, based on the clustering analysis method, through optimizing DBSCAN algorithm key blocks in catering, entertainment and trading industries were achieved direct extraction. The specific optimization was to use the non-location attributes in multi-dimensional check-in location data to calculate the weighted coefficient, modify the similarity function to determine the classification of data points, and construct the weighted density clustering algorithm to extract the high- heat blocks in urban industries. Based on the results of clustering extraction, the distribution characteristics of urban industries were summarized. The simulation results showed that the weighted DBSCAN of multi-dimensional spatial data could achieve the direct extraction of spatial characteristics of urban industries. The key blocks extracted had obvious characteristic with high popularity degree. The feature analysis of key blocks based on clustering extraction results was more reasonably reflect the development of urban industries. It would also provide an important reference for urban managers to make scientific decisions on industrial layout.