Neuro-rough models for modelling HIV

Abstract

This paper proposes a new neuro-rough model for modelling the risk of HIV from demographic data. The model is formulated using Bayesian framework and trained using Markov Chain Monte Carlo method and Metropolis criterion. When the model was tested to estimate the risk of HIV infection given the demographic data it was found to give the accuracy of 62% as opposed to 58% obtained from a Bayesian formulated rough set model trained using Markov chain Monte Carlo method and 62% obtained from a Bayesian formulated multi-layered perceptron (MLP) model trained using hybrid Monte. The proposed model is able to combine the accuracy of the Bayesian MLP model and the transparency of Bayesian rough set model.

Flexibly-bounded Rationality

“…Marivate and Marwala (2008) applied relational networks to successfully build a causal model that related the demographic characteristics to HIV status whereas Marwala and Crossingham (2008) successfully applied Neuro-rough technique to build a causal model that related demographic data to HIV status. Mistry et al (2008) successfully applied ensemble techniques to build a causal model that related demographic data to HIV status.…”