Abstract:

It is difficult to correctly forecast mobile telephone traffic data by the traditional method because of the strong nonlinearity of the data.In order to overcome the difficulty,this paper proposes a new approach to forecast mobile telephone traffic data.According to the characteristics of mobile telephone traffic,the mobile telephone traffic data is divided into two parts:the periodic traffic modeled by the first closest clustering method and the trend traffic modeled by the liner regression method.Hereby,an adaptive optimized fuzzy logical forecasting system is de’ signed,which is then tested by all data in a mobile telephone system of Guangdong Province.The testing results show that the proposed system iS effective in the forecasting of mobile telephone traffic.

Key words: traffic forecasting, first closest fuzzy clustering, traffic modeling