Resumen de Path construction of public management model based on POI social computing
Shujian Wen
In order to be able to better manage the production and life of urban residents, it is necessary to continuously optimizethe public management model, and this paper proposes the establishment of a public management path for POI socialcomputing. Build a computing system that can communicate with multiple computing units so that public managementgeneration and dissemination moves to the system boundary and users can communicate, share and collaborate using avariety of methods. Taking full advantage of the complete confidence of POI point data location attributes and timelydata update, puGAN model is added to improve the integrity of the collected data and distinguish the data sources bylearning the differences between real and pseudo samples. Generate data discriminative classification of real unlabeledsamples with unlabeled samples, adjust the distribution characteristics of the learned sample data, and improve thediscriminative ability. The gradient value of the sample discriminator is calculated, and the gradient generator is updatedto learn according to the data classification and finally solve the public management variance features. The analysis resultsshow that the puGAN model can improve the accuracy of POI localization, and the training error and testing error aremaintained at about 15%, which provides an important role for public management model research
