Forecasting of Traffic Load for 3G Networks Using Conventional Technique

The degrading performance of network coverage, resource allocation, and utilization is due to the rapidly increasing number of cellular subscribers, which is immensely difficult to predict the traffic load in Nigeria. The available developed algorithms and models did not well consider the behavior of the traffic load using the adopted input variables of this research. This paper arguably constructs an Artificial Neural Network (ANN) as Conventional technique to forecast the instantaneous traffic load per cell or eNodeB of 3G networks in Kano Metropolis. Four Active 3G networks data was extracted and recorded with the aid of W995 TEMS Pocket Phone over thirty five cells. The forecasted models when tested apparently tracked the measured traffic load with RMSE of 0.148365%, 0.21878%, 0.3327% and 1.32220%, thus achieved MAPE of 0.00394%, 0.00696%, 0.00109% and 0.03978% for A, B, C and D networks respectively. These validated that the Conventional technique can be a valuable tool in forecasting traffic load in Nigeria and could also be adopted in forecasting of large-scale metropolis cellular networks.