Human Health Assessment Based On Improved Probabilistic Neural Network

With the continuous improvement of people's living standards,more and more people begin to pay attention to their health problems.At the same time,the requirements for medical quality are getting higher and higher.Due to the improvement of people's living standards and unhealthy living habits,traditional health assessment methods have not been able to meet the human health monitoring assessment of most people.Therefore,this paper uses the intelligent algorithm-based telemedicine evaluation system to achieve monitoring and evaluation of human health,reducing the workload of medical staff and inspiring the innovation of modern medical management.In this paper,Probabilistic Neural Network(PNN)is selected as the basic model.Through the research of PNN method and the actual needs of this topic,PNN is improved to complete human health assessment.Whether it is the short training time,good performance or the need for a large number of training sample sets,PNN is very suitable for human health assessment.However,since all training sample sets directly constitute the number of PNN hidden layer neurons,when the training samples are too large,the PNN network structure is too complicated,which increases the difficulty of hardware implementation.In order to reduce run time and simplify the problem that the PNN is complicated due to the huge training samples,this paper quotes the characteristics of the radial basis neural network(RBFNN)mode layer neurons less than the total number of training samples,and introduces the RBFNN mode layer and output layer directly into the PNN.In the network topology.By extracting the respective advantages of RBFNN and PNN,a Radial Basis Probabilistic Neural Network(RBPNN)is formed.In addition,since the network performance of RBPNN depends largely on the value of the smoothing factor,in order to carry out human health assessment more accurately,it is proposed to use the artificial fish swarm algorithm(AFSA)to optimize the smoothing factor vector and dynamically adjust it.Eat behavior to increase the speed of computing.However,the health assessment of RBPNN after adjusting the smoothing factor by AFSA is easy to fall into the local optimum situation,which leads to unstable model results and cannot be directly used in human health assessment.In order to solve the problem that AFSA falls into local optimum in the smoothing factor optimization process,it is proposed to use the simulated annealing algorithm(SA)to improve the AFSA search later.The SA operation has the characteristics of global optimization.Therefore,by selecting the highest food concentration in the late stage of AFSA search and using SA operation to perform local search,local optimization is realized,and finally the approximate accurate extremum of the local optimal solution is obtained.Experiments show that the accuracy of the model is improved by optimizing the smoothing factor of RBPNN by AF-optimized AFSA.The research of this topic can help medical staff to manage human health more effectively,and also help individuals to understand their health status in a more real-time and comprehensive manner.The simulation results show that the accuracy and running time of the method are better than other methods.Therefore,the development of this topic can not only contribute to the modern medical management field,but also provide ideas for the improvement and application of PNN.