International Journal of Computational Intelligence Research (IJCIR)

Volume 3, Number 1 (2007)


Using inverse neural networks for HIV adaptive control

Brain Leke Betechuoh, Tshilidzi Marwala, Thando Tettey
School of Electrical and Information Engineering, University of Witwatersrand, Johannesburg, South Africa


Neural Networks are used in this paper, in an inverse configuration, for the adaptive control of HIV status of individuals. In the paper, a control mechanism, to understand how demographic properties (in this case, educational level) affect the risk of being HIV positive, is implemented. Preliminary design showed inverse neural networks outperforms the other methodology. The moral behind this implementation is to understand whether HIV susceptibility can be controlled by modifying some of the demographic properties such as education. The proposed method is tested on the HIV data set. It is found that the proposed method is able to predict the educational level of individuals to an accuracy of 88% if the HIV status of individuals and other demographic properties are known. It is thus possible to understand how the educational level of individuals can be modified to control the proneness of individuals to HIV contraction.

Multilayer perceptron, Feedforward neural networks, Inverse neural networks, Genetic algorithms, HIV, Adaptive control.