Research Of Medical Assistant Diagnostic Method Based On PCA-PSO-KELM Model

With the fast development of "Internet Plus" medical informatization,massive medical pathology data is generated every single day.The collection of pathologic data and simple processing approaches are familiar to the researchers.However,how to carry out data analysis and the extraction of knowledge from the massive data are still a new research field.Data Mining(DM)technology,which can be employed to analysis the big data,is playing an important role in the field of medicine.The classification algorithms,which can be used to predict the classification of unknown samples by studying and analyzing the training set of sample data,is one of the important components of the DM technology.In this thesis,the Artificial Neural Network(ANN)is used as the classification algorithm in data mining to obtain the correlation of pathologic data.Then a new pathologic data analysis model(a combination of PCA?PSO and KELM)is proposed,which is able to help physicians provide patients with a scientific and effective method of prevention measures.It should be noted that it is still in the development stage to apply the model to the research of assistant diagnosis of medical pathology data.The main achievements are presented as follows:1.Researching on the artificial neural network and existing correlative classification algorithms.The advantages and disadvantages of the algorithms are analyzed.This special attentions are paid on the basic principle and differences between the algorithms of the Support Vector Machine(SVM)and Extreme Learning Machine(ELM).2.The limitations of the two algorithms of SVM and ELM are discussed,a new model algorithm that is proposed to release the limitations of SVM and ELM.The new algorithm introduces the kernel concept of SVM into the framework of ELM model,and forms the Kernel Extreme Learning Machine(KELM).In order to get stable system parameters for KELM,an efficient Particle Swarm Optimization(PSO)is used to optimize the penalty parameters and kernel function parameters.The PSO parameter optimization approach finally leads to the PSO-KELM model.3.In order to reduce the input dimension and accelerate the convergence speed of the following analysis model,the Principal Components Analysis(PCA)algorithm is employed.Combining with the PCA and PSO-KELM and the medical data of breast cancer and coronary heart disease were analyzed based on the simulation of experimental medical data.The experimental results of the model are compared with the results of Learning Vector Quantization(LVQ),SVM and KELM model.After the simulation,we know that this new algorithm has the attribution of fast convergence speed and high accuracy performance.