Page 95 - ITU Journal - ICT Discoveries - Volume 1, No. 2, December 2018 - Second special issue on Data for Good

ITU JOURNAL: ICT Discoveries, Vol. 1(2), December 2018

APPLICATION OF CANFIS MODEL IN THE PREDICTION OF MULTIPLE-INPUT
TELECOMMUNICATION NETWORK TRAFFIC

Francis Kwabena Oduro-Gyimah , Kwame Osei Boateng
1
2
1 Ghana Technology University College, Department of Telecommunications Engineering, Kumasi, Ghana,
1, 2 Kwame Nkrumah University of Science and Technology, Department of Computer Engineering, Kumasi, Ghana

Abstract –Telecommunication network traffic prediction is an important approach that ensure efficient
network planning and management. Telecommunication network traffic is univariate and prediction
models have mostly been concentrated on single-input and single-output traffic. This study proposes a new
approach, the multiple-input multiple-output Coactive Neuro-Fuzzy Inference System (CANFIS) model to
predict a five time span univariate hourly, daily, weekly, monthly and quarterly time series of 3G downlink
traffic simultaneously. In the modelling process several parameters were used in the configuration of the
network. The best model for predicting five-input telecommunication traffic was CANFIS (5-2-5) which
employed a Bell membership function, Axon transfer function and Momentum learning rule and the
membership function per input of 2. The performance of the model was evaluated by comparing the
predicted traffic with actual traffic obtained from a 3G network operator and the results indicate a
minimum accuracy measure value of MSE = 0.000486, NRMSE = 0.01120 and percent error = 12.33%.

Keywords – 3G downlink, CANFIS, multiple-input, multiple-output, prediction, telecommunication
network traffic.

1. INTRODUCTION Several methods have been used to develop high
precision techniques in forecasting 3G network
Telecommunication network traffic prediction is an traffic [9] [10] [11]. Reference [10] applied data
important approach that ensures efficient network mining technique in predicting the air interface load
planning and management. However, in research, of 3G network traffic while reference [12]
the focus of forecasting mobile network traffic has established that 3G cellular network resource
mostly been on developing single models for each management is influenced by factors such as
individual data set [1] [2][3][4]. Other research number of users, multipath propagation, congestion
works that have applied individual traditional control, transport protocol flow, etc.
forecasting approaches for prediction are Kalman
filtering [5], ARIMA and exponential smoothing [6] Reference [13] designed a multiple fuzzy system
and voice traffic forecasting for GSM using architecture which is connected side by side. The
feedforward neural network [7]. Reference [8] model is able to predict time series data at
applied four different models, linear regression, dissimilar inserting lengths and time intervals;
simple exponential regression, ARMA and dynamic however the system could not predict
harmonic regression (DHR) to analyze hourly, daily instantaneous multiple outputs.
and weekly telecommunication traffic. Reference [9]
applied a neural network ensemble to 44160 hourly In another study, [14] proposed a multiple-input
data of HSDPA traffic and indicated that the neural multiple-output Adaptive Neuro Fuzzy Inference
network ensemble predict the traffic with high System (MANFIS) model that considered overtaking
accuracy. incidents of vehicles for dissimilar time steps in the
future. This model however, predicted two different
time steps of the future as output using five inputs.
However to the best of our knowledge the Coactive

© International Telecommunication Union, 2018 73