Research On Intelligent Grading Method For Diabetes Nursing

The rapid development of information technology and the change of medical models have brought diabetes nursing into a new stage.The exploration of effective diabetes nursing models and related technologies have become a hot spot in the academic and medical circles.In view of the problems of single nursing model and poor generality of research methods in traditional diabetes nursing research,this thesis combines machine learning and diabetes nursing to study a diabetes nursing grading model based on machine learning.This model is effective to form an early warning mechanism for possible physical and mental health problems.This mechanism to prevent the deterioration of the patient’s condition,reduce the burden on individuals and medical insurance and achieve scientific nursing for diabetes.This thesis takes diabetes nursing as the research object and studies the classification method of diabetic patients in the nursing process.The main tasks are as follows:This thesis constructs a hierarchical model for diabetic physiological care.Aiming at the problem of physiological nursing in diabetes care,this thesis proposes a grading model of physiological nursing based on improved adaptive decision tree.This model analyzes the relationship of the attributes of the data set,combines feature reduction,feature fuzzification and Tikhonov filtering technology for preprocessing the data,and selects the split attribute using ID3 and C4.5 when classifying the decision tree.The improved adaptive decision tree is constructed from the pre-processed data and then the rules are extracted.Then the diabetic patients are divided into different levels to form a diabetic physiological nursing classification method.The experiment proves the feasibility of algorithm based on improved adaptive decision tree and the effectiveness of Tikhonov filter technology.At the same time,the accuracy,recall and precision of different algorithms are compared through experiments,and the experiment proves that the improved adaptive decision tree algorithm has higher accuracy,recall and precision than adaptive decision tree,decision tree,improved support vector machine,support vector machine.This thesis constructs a hierarchical model for diabetic psychological care.Aiming at the psychological nursing problems in diabetes nursing,this thesis proposes a psychological nursing grading model based on pigeon-flock K-Means which clusters patients based on similar mental states and then classifies them based on mental health scores.This model analyzes the characteristics of the questionnaire data set,uses word2vec to vectorize text data,and combines TF-IDF to obtain the text of each mental health questionnaire vector.According to the Euclidean distance between the vectors,the original data is reduced in dimensionality.So the parameters and complexity of the model are reduced.After data processing,this thesis constructs a mental nursing grading model that only considers the patient’s mental health cognition,and a mental nursing grading model that considers the cognitive deviation of patients’ mental health between caregivers and patients.The experiment shows the feasibility of the two psychological nursing grading models,and the determination of the psychological grading level.Simultaneously,the simulation experiment shows that the efficiency of pigeon-flock K-Means is higher than particle-swarm K-Means algorithm and K-Means algorithm.