Investigation of effect of number of hidden neurons in Statistical Neural Network

One of the most important characteristics of a perception network is the number of neurons in the hidden layer(s) and researchers have had a problem in determining the number of hidden units to obtain optimal network performance. This study employed Statistical Neural Network (SNN) using hidden neuron in predicting mother to child transmission of HIV. The data obtained from ANC of a certain hospital and simulated data were used for the study. Mean Square Error (MSE), Akaike Information Criteria (AIC) and Neural Information Criteria (NIC) were computed for determining the adequacy of the hidden neuron that determines the optimality of the model. The best result shows that the hidden neuron that determines the optimality of the model in the prediction of HIV with CD4 as an input variable with 7 neurons is SNN (1-2-1).

Abdalla, S. W. (2011). Mathematical modelling of HIV/AIDS dynamics with treatment and vertical transmission. M.Sc. Dissertation, University of Dar es Salaam.

Abraham, T. (2005). Application of data mining technology to identify determinant risk factors of HIV infection and to find their association rules: the case of Center for Disease Control and Prevention (CDC). Master's Thesis, Addis Ababa University, Addis Ababa.

Bishop C. M. (1995). Neural Networks for Pattern Recognition. Oxford: Oxford University Press

Ganesan, N., Venkatesh, K. and Palan, A. M. (2010). Application of neural networks in diagnosing cancer disease using demographic data. International Journal of Computer Application: 1-14.

Geman, S., Bienenstock, E. and Doursat, R. (1992). Neural networks and the bias/variance dilemma. Neural Computation, 4(1), 1-58.

Gnana, S. K. and Deepa, S. N., (2013). Review on methods to fix number of hidden neurons in neural networks. Journal of Mathematical Problems in Engineering. doi.org/10.1155/425740

Shafi, I., Ahmad, J., Shah, S. I., and Kashif, F. M. (2017). Impact of varying neurons and hidden layers in neural network architecture for a time-frequency application.www.researchgate.net/publication/224703002

Jeff, H. (2009). Introduction to Neural Networks. MIT Press, Cambridge, MA.

Keeni, K., Nakayama, K. and Shimodaira, H. (1999). Estimation of initial weights and hidden units for fast learning of multi-layer neural networks for pattern classification.Proceedings of the International Joint Conference on Neural Networks, 3: 1652-1656.

Kwak, N. K. and Lee, C.(1997). A neural network application to classification of health status of HIV/AIDS patient. Journal of Medical Systems, 21(2).

Lee C.W. and Park J.A. (2000). Assessment of HIV/AIDS-related health performance using an artificial neural network. Information & Management, 38(4) 231-238.

Onoda, T. (1995). Neural network information criterion for the optimal number of hidden units. Proceedings of the 1995 IEEE International Conference on Neural Networks,1: 275-280.

Rumelhart, D. E, Hinton, G.E. and McClelland, J.L. (eds) (1986). Learning internal representations by error propagation. In: Parallel Distributed Processing: Explorations in the Microstructure of Cognition, MIT Press, Cambridge, pp: 318-362.

Sibanda, W. and Pretorius, P., (2011). Novel application of multi-layer perceptrons neural networks to model HIV in South Africa using seroprevalence data from antenatal clinics. International Journal of Computer Applications, 35 (5), 26-31.

Udomboso, C. G., James, T. O. and Mba, O. O. (2012). On R2 contribution and statistical inference of the change in the hidden and input units of the statistical neural networks. Journal of the Nigeria Association of Mathematical Physics, 22.

UNAIDS (2010). Report on the global AIDS epidemic 2010. Geneva: Joint United Nations Programme on HIV/AIDS.

WHO (2009). AIDS epidemic update.

Wu, Y. K. and Hong, J. S. (2007). A literature review of wind forecasting technology in the world. Proceedings of the IEEE Lausanne Power Tech, pp. 504509.

Xu, S. and Chen, L. (2008). A novel approach for determining the optimal number of hidden layer neurons for FNN's and its application in data mining. Proceedings of the 5th International Conference on Information Technology and Applications, pp. 683-686