Fine prediction method of mobile communication network traffic based on user behavior mining

Abstract

The invention discloses a fine prediction method of mobile communication network traffic based on user behavior mining, and the method comprises the steps: collecting the data of the core network sideof a mobile communication network, and forming complete Internet record log information for the collected mobility management data and Internet data according to each service event; establishing a multi-dimensional cell characteristic model including a user resident characteristic, a user movement characteristic, a time period peak shifting characteristic and a service development characteristicbased on the Internet record log information of the user, and counting basic data of each cell; performing random forest training classification on the basic data, and summarizing the cells into different classes according to user laws, business laws and load laws; and carrying out different time sequence prediction model training on different types of cells, carrying out continuous calibration, and establishing a service traffic prediction model system close to reality. The method provided by the invention effectively solves a series of problems of difficult prediction, long prediction period, large prediction and actual errors and the like in the past.

Description

technical field [0001] The invention belongs to the field of combining machine learning and big data, and in particular relates to a method for refined prediction of mobile communication network traffic based on user behavior mining. Background technique [0002] Traditional business traffic forecasting usually simply adopts linear programming algorithms, but the traffic is not a simple linear growth, but presents periodic changes, and the forecast results deviate greatly from the actual, especially during major events or holidays, traffic data There will be a sudden increase. According to the previous general linear prediction model, it is impossible to predict the sudden increase of festival traffic, so the prediction error of the traffic is relatively large, which cannot meet the actual needs of users and operators. [0003] With the background of the transformation of the network operation work from the previous rough style to the refined one, the traffic forecasting wor...

AI Technical Summary

Problems solved by technology

[0004] The purpose of the present invention is to provide a method for refined forecasting of mobile communication network traffic based on user behavior mining, so as to realize more accurate forecasting of mobile communication network traffic, thereby solving the difficulties in previous forecasting, long forecasting period, forecast and actual error Larger and other series of problems provide strong data support for resource assessment, deployment, and avoiding large traffic impacts during holidays and important events

Images