Research And Implementation Of Intelligent Disease Diagnosis And Prediction System

With the rapid development of computer technology,the medical field has begun to use these technology to improve the level of medical services.In addition,the rapid development of medical information technology has also accumulated a lot of medical data in the medical information system.Deep learning algorithms can learn abstract features from data.These features can be used to identify and classify the targets.How to apply deep learning methods to medical data and establish analytical model is a hot topic in the current medical data analysis field.The thesis mainly studies the application of deep learning theory in the medical field.We designed and trained different deep learning models based on different types of medical data and task requirements.At the same time,the prototype of the disease prediction and diagnosis system was implemented.The main work is as follows:First,we explored the classification and prediction application of convolutional neural networks on a one-dimensional structured newborn birth defect data set.Birth defects are abnormalities of body structure,function or metabolism that occur during the fetal period,which seriously affect the survival and quality of the newborns.At present,there are only some preventive measures to deal with birth defects in our country.There are not many studies on the causes of birth defects in fetuses.A one-dimension convolutional neural network is designed and trained for one-dimension structured birth defect data.Also,a model used to classify the four common birth defect diseases is established.The experimental results show that the prediction model has a high prediction accuracy rate and has certain practical value.Secondly,for two-dimensional fundus images,we explored the application of convolutional neural networks in multi-class and multi-label classification problems.The diagnosis of eye diseases mainly relies on fundus images.At present,many results have been achieved in the research of classification models for the diagnosis of single eye diseases.However,the research on classification models of patients suffering from multiple eye diseases is still in initial stage.Therefore,we designed an eye disease diagnosis model based on VGG network.The model combines the age and gender information module and the attention mechanism module.The results show that the model has a high classification accuracy for several common eye diseases.The addition of the age and gender information module and the attention mechanism module also significantly improves the classification accuracy of the model.Based on the results above,we designed and implemented the prototype of the intelligent disease diagnosis system.The system is developed based on the B/S architecture and includes two subsystems: birth defect prediction and eye disease diagnosis.The user can log in the system main page to operate through the browser on the mobile phone or computer.The interface of the entire intelligent disease diagnosis system is simple,easy to operate,and has a certain application and promotion prospect.