Research Of Intelligent Diagnosis For Short Children Based On Neural Network Optimized By Genetic Algorithm

Short stature is of common pediatric diseases in impeding the development ofchildren ’s physical and mental health. It is not only the patient’s personal problems butalso social problems what had to face. Short children ’s screening, diagnosis, treatmentis a complex clinical diagnostic technology, even experienced clinical pediatrician oftenmisdiagnosis due to the difference of professional background, professional orientationand clinical experience. Therefore, combining the evaluation of advanced sciencetechnology with extensive clinical experience, using advanced information technology,achieving computer-aided diagnosis of short stature, not only can improve the clinicalscreening, diagnostic accuracy and labor efficiency, but also allows the majority ofpatients diagnosed and treated at the basic level area timely and effectively.Because of the problem of the short children, an expert system based on neuralnetwork be established by using a mode of computer-aided diagnosis. A new neuralnetwork model is proposed which is based on optimized BP(Back Propagation)neuralnetwork optimized by genetic algorithm. The superiority with the model of BP neuralnetwork and genetic algorithm were fully integrated.The index system of diagnosing the short children was brought forward. Then combineGenetic Algorithm with BP neural network, using GA’ s global search capability the BPnetwork structure parameters are optimized. Finally, the experiment results show thatnew algorithm can overcome the local convergence, feebleness of generalization abilityand other issues of BP algorithm effectively. The model works with quickerconvergence rate and the higher precision, and is suitable for the computer-aideddiagnosis of short children.