Network flow prediction method based on attention multi-component space-time cross-domain neural network model

Abstract

The invention discloses a network traffic prediction method based on an attention multi-component space-time cross-domain neural network model, belongs to the technical field of intelligent communication, and solves the problem of prediction of wireless cellular network traffic. The method comprises the following steps: dividing wireless cellular flow data into neighbor data, daily period data andweek period data according to periodic characteristics; modeling the neighbor data, the daily period data and the week period data through a conv-LSTM structure or a conv-GRU structure; distributingdifferent weights to the three kinds of feature data in a self-adaptive mode through an action layer, improving the feature extraction capacity of the three kinds of feature data, and restraining feature information interfering with the prediction moment; and finally, in combination with timestamp feature embedding and multi-cross-domain data fusion, jointly assisting the model to perform trafficprediction. The model can effectively utilize the periodic characteristics of the wireless cellular traffic data, saves the model training time, greatly reduces the workload, and further improves theprediction performance of the network traffic.

Description

technical field [0001] The invention belongs to the technical field of intelligent communication, and in particular relates to a network traffic prediction method based on an attentional multi-component spatiotemporal cross-domain neural network model. Background technique [0002] With the advent of the 5G / B5G era, the number of mobile devices and the Internet of Things is growing exponentially around the world, and people's demand for wireless mobile data is growing rapidly. How to scientifically and rationally allocate and optimize existing cellular network resources, improve resource utilization, and reduce energy consumption of cellular base stations is a problem that the communication industry needs to think about and solve. [0003] Accurate forecasting of wireless cellular traffic is helpful for base station site selection, urban area planning, and regional traffic forecasting. However, accurate prediction of wireless service traffic is a very challenging problem, m...

AI Technical Summary

Problems solved by technology

Second, the spatial constraints caused by multi-source cross-domain data on wireless service traffic

Third, how to achieve high prediction accuracy of wireless cellular traffic while considering spatio-temporal factors and combining cross-domain data is also a difficult problem

Images