ARMA-ARCH model family based prediction method for GPRS (general packet radio service) data services of GSM (global system for mobile communications) communication system

Abstract

The invention discloses an ARMA-ARCH model family based prediction method for GPRS (general packet radio service) data services of a GSM (global system for mobile communications) communication system. The prediction method includes the key steps: preprocessing the GPRS data services by seasonal period determination, ln downsizing, deseasonalizing and detrend; and using an ARMA-ARCH model family for prediction of preprocessed GPRS data. By the prediction method, the GPRS data services can be effectively predicted, an important reference means is provided for operators to describe current network operating statuses and perform performance parameter analysis and planned capacity expansion, maintenance and optimization of a wireless communication network during communication, data traffic change tendency of the network can be predicted in advance, capacity expansion and optimization of the network can be achieved in a planned way, and considerable economic benefit is created directly or indirectly.

Description

Technical field: [0001] The invention relates to GSM communication, especially the technical field of prediction of GPRS data service flow in the communication system. Background technique: [0002] Looking at the development trend of the wireless mobile communication market, an obvious feature in recent years is that users' demand for mobile data services continues to rise. Especially in recent years, with the intensification of competition in the mobile communication market, the tariffs of mobile traffic services have shown a downward trend. In stark contrast to the continuous decline in revenue of traffic services, data services are forming a new revenue growth point, and this The new source of income will also maintain a continuous growth trend. Although the rapid growth of data services has brought unlimited business opportunities, it has also brought new challenges to mobile communication operators' network capacity and optimization. GPRS network is currently the mai...

AI Technical Summary

Problems solved by technology

[0012] Judging from the current patents or related academic research, there is no relevant research that predicts data services from the perspective of volatility, which greatly wastes the amount of information contained in the data. From another perspective, the information of the data is not fully mined. The understanding of the data is not clear enough, and the best prediction effect will inevitably not appear; the relevant research has not performed the necessary preprocessing on the data before using the ARCH model to extract the long-term autocorrelation of the data itself and the application of the ARMA model to extract the data itself. Steps for short-term autocorrelation

Images