Research On Intelligent Medical Diagnosis Auxiliary Method Based On Machine Learning

With the development of intelligent computing technology,the intelligent assistant system is used world widely.In modern medicine,doctors use their personal experience and medical knowledge to diagnose the disease,and draw conclusions.In order to effectively inherit the medical diagnosis experience accumulated by doctors,researchers have proposed the idea of using artificial intelligence technology to develop intelligent medical diagnosis assistant system.Aiming at the above problems,the machine learning technology is introduced into the medical field.In this paper,we discuss a new method of intelligent medical diagnosis based on machine learning,and design and implement the intelligent diagnosis assistant method by using the existing medical record data.The main research work of this paper includes the following parts:First of all,based on the application technology of machine learning in natural language processing,this paper designs the framework of intelligent medical diagnosis assistant system for electronic medical records.The diagnostic assistance problems are summarized as data processing and medical records classification.The data processing part.For the medical data of medical records has a obvious medical professional language semi-structured,multi-dimensional information and other characteristics.In this paper,we design a semantic propagation algorithm to complete the medical language feature based on the medical language lexicon to establish and create a medical language characteristics of the text matching model.So as to realize the vectorization of electronic medical records and the feature extraction based on matching.In the case of text classification,text oriented electronic medical record specific data,we use the typical support vector machine(SVM)and multilayer perceptron(MLP)two kinds of classification methods,and the classification of diseases are,and then through the evaluation of the experimental results,and compared the performance of two kinds of methods,evaluation conclusion that MLP is better than SVM,which provides the reference for the development of practical application system.Furthermore,according to the popular technology of LSTM in deep learning,the corresponding experimental results are obtained by processing the same electronic medical record data.By comparison with SVM and MLP two technical results,we found that in the analysis of the electronic medical record text oriented small amount of data in the LSTM model of deep learning intelligent computing ability can not be brought into full play,the MLP classification method has the best performance in traditional machine learning.