Research On Medical Health Classification Based On Machine Learning

With the continuous improvement of people’s living standards,physical health has become more and more concerned.Physical examination has become an important way for people to pay attention to their own health.In the face of disorganized physical examination data,this article uses machine learning methods to processing the medical data,and uses the processed data to train the model,and finally obtains the accurate classification result of the disease.This article gives a brief analysis of the classification methods on medical health data sets.Firstly,the concept of machine learning is introduced.The classification of machine learning algorithms is introduced from five aspects: supervised learning,semi-supervised learning,non-supervised learning,reinforcement learning and genetic algorithm.Second,artificial neural network and biological neuroscience,health archives and electronics are introduced.The relationship between the medical records;Finally,a brief analysis of the classification algorithms in machine learning was conducted.The main research work of this paper is divided into the following two aspects:First,the performance comparison of the machine learning classification algorithm on the public medical data set.In this paper,the experimental environment and experimental tools used are introduced.From the introduced classification algorithms,support vector machines,decision trees,deep neural networks and convolutional neural network classifiers are selected in the UCI machine learning resources Statlog,Audiology,Adult and Heart.Disease Public Medical Database conducts classification experiments and conducts comprehensive comparative analysis.The results show that the convolutional neural network classifiers in the four classifiers achieve better classification results compared to the other three classifiers in the classification of public medical data sets.Second,classification studies are performed on medical datasets for convolutional neural network classifier performance.This article analyzes the characteristics of medical data sets and makes necessary data cleansing for the characteristics of medical data.We introduce the process of constructing a convolutional neural network model in detail,including the selection of activation functions and the initialization of weights in neural networks.The cleaned data was trained on a constructed convolutional neural network.After many experiments,the accuracy of the classification of the convolution neural network classifier selected in this paper on the medical examination data set was 97.5%.In summary,in the research of medical examination data,the convolutional neural network used in this paper has achieved good classification effect in the classification of diseases with large amounts of data and multiple categories.With the continuous advancement of the hospital’s informatization,medical data will gradually be standardized;people will continue to deepen their understanding of machine learning and believe that more accurate classifiers will emerge.