Research On Prediction Method And System Design Of Cerebral Infarction Based On Machine Learning

The incidence,disability and mortality of cerebral infarction are high and the prognosis is poor,which seriously endangers human health.However,studies show that giving scientific and reasonable comprehensive nursing intervention in the early stage of the disease can promote the patients to recover faster and better.Therefore,the accurate prediction of cerebral infarction is particularly important.Machine learning,as a discipline that studies how to learn and mine information from data,has been widely used in the field of disease prediction,showing unique advantages such as accuracy,sensitivity and generality.Therefore,based on the establishment of cerebral infarction data set,this paper builds a prediction model of cerebral infarction based on machine learning and designs a cerebral infarction health management system.The main research contents and innovations are as follows:(1)In order to improve the accuracy and generalization ability of the prediction model,based on the stacking selective ensemble learning,the selection method of the base learner combination is optimized and improved,and a prediction model of cerebral infarction based on the AD-Stacking ensemble framework is proposed,the model comprehensively considers the accuracy and difference of the base learner,and uses the harmonic mean of accuracy and difference as the basis for selecting the combination of base learners,which not only solves the problem of insufficient diversity of base learners due to only considering the accuracy,but also solves the problem of mixing low precision base learners caused by only considering difference,effectively improving the accuracy of cerebral infarction prediction.(2)In order to explore the performance of deep learning in the predictive modeling task of non-sequential structured data,an attention mechanism-based prediction model for cerebral infarction ADFM is proposed.On the one hand,the low-and high-order combined features are fully extracted through the LR module,the second-order feature intersection module and the DNN module;on the other hand,the attention mechanism is introduced to adaptively learn the importance of each sub-module,and assign higher values to the key modules,so as to make full use of key information and suppress useless information,thereby improving the generalization performance of the model.ADFM model provides a new solution for cerebral infarction prediction.(3)In order to alleviate the problems of uneven distribution of medical resources and imbalance of doctor-patient ratio,a cerebral infarction health management system based on Django framework was designed.The system includes functions such as account management,user information management and cerebral infarction prediction,realizing the dynamic storage and cross-access of medical data,and providing an intelligent auxiliary diagnosis platform for cerebral infarction for users.The two proposed prediction models are verified on the cerebral infarction dataset constructed in this paper.Compared with the basic model,the accuracy are increased by 1.41%and 2.27%,respectively.At the same time,the realization of the health management system provides decision support for the prevention and treatment of cerebral infarction.The prediction method research and system design of cerebral infarction based on machine learning have certain reference value for the prevention and management of cerebral infarction.