Application Of Machine Learning In The Diagnosis Of Internal Diseases

The diagnosis of internal medicine diseases depends on doctors' clinical experience to a great extent,which they could broaden through extensive clinical practice and literature reading,including generalization and summarization of previous cases.Meanwhile,there are increasing auxiliary diagnosis and treatments to enhance accuracy and efficiency of diagnosis,and machine learning is one that offers excellent potential to so.A large number of preceding cases can be analyzed and then some patterns of diseases can be concluded via this state-of-the-art technique-machine learning,leading to the jump of diagnosing efficiency and enabling doctors to have more time for adjuvant therapy.Hence,applying machine learning and natural language processing into diagnosis deserves significant attention.This paper first introduces the background and significance of machine learning and medicine.Then there is pretreatment about the information from 371 patients who were admitted to hospitals for cough and got final diagnoses.Pretreatment includes reading the case text file with Python,extracting the case eigenvalues and processing them mathematically.The data about 269 patients from the first hospital are used as a training set,and the other data about 102 patients from the second hospital are treated as a test set.In the training set,algorithms including Logistic regression,Support vector machine and Random forest which are used to construct prediction models.In the test set,the performances of the models from the aspects of accuracy,precision,recall,F1-score are evaluated.The results showed the Pros and Cons of the different algorithms.Finally,combined with the clinical problems and the application of the models,the ultimate outcomes were evaluated.In this paper,patients admitted to hospitals for treatment due to cough are the main subjects of study.In essence,this paper studies the multi classification of machine learning.The theory,model and algorithm mentioned in this paper have certain generalization,and have certain guidance and reference significance for case reading and other medical diseases auxiliary diagnosis.