Spatiotemporal data modeling and prediction algorithms in intelligent management systems

In order to solve the problem of difficulty in learning semantic pattern representations between user dynamic interest sequences using path based and knowledge graph based entity embedding methods, the author proposes research on spatiotemporal data modeling and prediction algorithms in intelligent management systems. The author first makes a preliminary analysis of the wireless network data (mainly the data of cellular mobile networks) obtained by Internet service providers, reveals that the data of adjacent base stations have temporal and spatial correlations, then establishes a hybrid deep learning model for spatio-temporal prediction, and proposes a new spatial model training algorithm. Finally, experiments were conducted using wireless network datasets to evaluate the performance of the model. The experimental results show that based on data analysis, it can be seen that the prediction of the system has effectively improved by 99%. The spatiotemporal data modeling and prediction algorithm proposed by the author in the intelligent management system significantly improves prediction accuracy.