Neonatal Disease Diagnosis : AI Based Neuro-Genetic Hybrid Approach

Accurate disease diagnosis and proper management is a matter of concern to everybody. In this context, there are a number of uncertainties involved including human errors. The problem is augmented when the domain is itself critical; for example, neonatal diseases. The problem is further augmented whenever and wherever proper experts are not available. Mitigating the kind of such problems, developing medical decision support systems using AI techniques are being explored during last few years. Combining neural networks with genetic algorithms is one of the approaches people use. It reduces various medical errors and provides better prediction of diseases. This article presents a study of neuro-genetic approach with multi layer perceptron (MLP) neural network diagnosing neonatal diseases. Integration of genetic algorithm and incremental back propagation neural network has been found suitable for a better trained network; experimental results and a comparative study corroborate the fact.