The Research And Design Of The Intelligent Method Of Auxiliary Diagnosis Of Asthma

In recent years,intelligent technology development,to promote the application of intelligent auxiliary system,in the modern medical doctors to use its rich work experience and reserves to diagnosis and treatment of patients,medical knowledge in artificial intelligence technology,more and more applied to medicine to help doctors auxiliary diagnosis,this paper introduce the machine learning technology in the field of Chinese medicine asthma,discussed a kind of intelligent diagnosis based on optimization of BP neural network,the auxiliary method,and use the text and image of the existing hospital medical record,design the framework of traditional Chinese medicine(TCM)for medical record the wisdom of medical,the auxiliary diagnosis problems were summarized as data processing and diseases classification,mainly including the processing of the picture data of the palm lines of the main symptoms of asthma and the processing of the text data of TCM medical plans.The processed data were used to classify the symptoms and syndrome types of asthma.The main work contents were as follows:(1)It has been proved by Chinese medicine for many years that there is a direct correlation between the type of asthma and the Yin and Yang characteristics of the palm lines.In this paper,we use the method of deep learning to train the original data set of large palm lines in the hospital,this paper introduces the specific process and advantages of CornerNet-Saccade algorithm,this paper introduces a network of delocalization Non_Local using Non_Local network easy portability and focus on the characteristics of global optimize the CornerNet-Saccade algorithm,In this paper,a simple and easy to use data set labeling tool-LabelImg is introduced.After labeling the data set,the data is enhanced by off-line.Finally,the optimized cornernet-saccade algorithm was used to identify and classify the palm lines of big fish.The experimental results show that the optimized algorithm has a great improvement in the recognition and classification of big fish palm lines.(2)In order to achieve the objective description of text medical record data,this paper made the pretreatment of the text,and wash out has nothing to do with the disease syndrome and symptoms of text,and then use stutter segmentation tools,the text of the medical records after cleaning for word segmentation,text after the word again using TF-IDF keyword extraction,extract the key asthma syndrome,will process after the big thenar palmprint recognition results added to the basis,and then through the map function layer upon layer traversal to quantify the critical asthma syndrome,finally to be nice to quantitative data with the method of normalized processing,eliminate the differences caused by different dimensions.(3)According to the applicable objects and characteristics of the subject,by adding way of vector and to fall into local minimum to shake up the model,on the other hand from the vector,the number of hidden layer nodes,and optimized the number of iterations,overcomes the traditional BP neural network easy to fall into a fitting phenomenon and the most superior local defects,through the use of simulation analysis and comparison with other algorithms,to verify the validity and accuracy of the model,and improves the accuracy of classification of traditional Chinese medicine for asthma.