OPEN ACCESS
The aim of this work is to model the pathway from caller to recipient of GSM telecommunication in Nigeria, with a view to produce a model that can help reduce the problem of drop calls experienced in the industry. Our dependent variable was total successful calls against 9 explanatory variables. An initial multiple linear regression produced low R2 of 25.5%. Diagnostics interventions of some transformations with removal of leverage points improved the R2 to 82%. Two model selection techniques, Mallow’s Cp and adjusted R2 were used to obtain the best parsimonious model, which contained 7 explanatory variables. The results show that the main variables that explain total successful call are Percentage drop calls, Proportion of transmission failure, Call traffic Congestion, Control channel failure, Earlang, P_HR and Availability. We, therefore, advise telecommunication industries in Nigeria to use the model to counteract the problem of drop calls.
Telecommunication pathway, drop calls, Box-Cox transformation, Mallow’s Cp, model selection.
[1] I.A. Adeleke, E.E.E. Akarawak, A.I. Opara, E.O. Esan. “Box-Cox Transformation in Regression Analysis: An Illustration with Telecommunication Data” Proceedings of the Nigerian Statistical Association, Asaba, Nigeria; 2007.
[2] A. Agrawal, M. F. Qureshi, “Indian Weather Forecasting using ANFIS and ARIMA based Interval Type-2 Fuzzy Logic Model”; AMSE Journals, Series Advances D, Vol. 19, no 1; pp 52-70, 2014.
[3] G.E.P. Box, W.G. Hunter, J.S. Hunter. Statistics for Experimenters, John Wiley & Sons,1978.
[4] L. Breiman. “Submodel Selection and Evaluation in Regression – The Conditional Case and Little Bootstrap”, Technical Report 169, Department of Statistics, University of California, 1988.
[5] T. Erven. Grunwald, P. & Rooij, S.; “Catching Up Faster in Bayesian Model Selection and Model Averaging”, 2009;. http://books.nips.cc/papers/files/nips20/NIPS2007_0756.pdf. Assess date September 2009.
[6] I. Fadeyibi. “Industry Analysis, Nigerian Mobile Telco, 2009; PRLog – Global Press Release Distribution”. www.prlog.org/10242511- industry-analysis-nigeria-mobile-telco.pdf. Assess date September 2009.
[7] E.I. George. “The Variable Selection Problem”; Journal of the American Statistical Association, vol 95, No 452, pp 1304-1308, 2000.
[8] F.A. Graybil. Theory and application of Linear Model; Belmat CA: Weadsworth, 1976.
[9] David C. Hoaghin & Roy E. Welsch; “The Hat Matrix in Regression and ANOVA”; The American Statistician, vol 32, no 1, pp 17-22, 1978. [10] Hurvich, C.M & Tsai, C.; “The Impact of Model Selection on Inference in Linear Regression”; The American Statistician. Vol 44, no 3, pp 214-217, 1990. [11] Iwundu, M.P. and Efezino, O.P.; “On the Adequacy of Variable Selection Techniques on Model Building”, Asian Journal of Mathematics & Statistics, vol 8: pp 19-34, 2015.
[12] Kadane, J.B. & Lazar, N.A.; “Methods and Criteria for Model Selection” Journal of the American Statistical Association. Vol 99, no 465. Review Article, 2004.
[13] Kennard, R.W.; “A Note on the Cp Statistic”; Technometrics, vol 13, no 4, pp 899-900 , 1971.
[14] Kundu, D. & Murali, G.; “Model Selection in Linear Regression”, Elsevier: Computational Statistics and Data Analysis. vol. 22, pp 461-469, 1996. [15] Kwon, Y.; Bozdogan, H. & Bensmail, H.; “Performance of Model Selection Criteria in Bayesian Threhhold VAR (TVAR) Models”; Econometric Reviews. Vol. 28, Issue 1-3, pp 83-101, 2009. [16] Lee, M. A.; “Application of model selection techniques and measures of agreement to advertising data”, an Unpublished Masters Dissertation, Dept. of Mathematics and Statistics, University of New Mexico; http://repository.unm.edu/handle/1928/32963; 2016.
[17] Mallows, C.L.; “Some Comments on Cp”; Technometrics, vol 15, no 4, pp 661-675; 1973.
[18] M.Y. Omotoso. Econometrics: A Practical Approach; Yosode Book Publishers, Ibadan; 2000.
[19] P.F. Li. “Box-Cox Transformation: An Overview”, Department of Statistics, University of Connecticut, www.stat.ucom.edu/~studentjournal/index; 2005. Assess date August 2007.
[20] R.P. Schoukens, G. Vandersteen, “Model Selection through a Statistical Analysis of Global Minimum of a Weighted Nonlinear Least Squares Cost Function”; IEEE Transactions on Signal Processing, vol. 45, no 3, pp 686-693; 1997.
[21] A.C. Rencher, F.C. Pun, “Inflation of R2 in Best Subset Regression”; Technometrics, 22, no 1, pp 49-53, 1980.
[22] S. Weisberg. Applied Linear Regression (2nd ed); Wiley, New York, 1985.