Research On The Correlation Model Of Frailty And Disease Based On Unsupervised Learning

Frailty is a kind of state or a group of syndromes when the body's multi system structure and multi organ tissue reserve function is reduced to the threshold value due to the defects in physical function,mental spirit and social security.Frailty has the characteristics of high complication and high risk,which will lead to the prolongation of diagnosis and treatment cycle,the reduction of cure rate,and the increase of mortality.The existing clinical medical research shows that the disease will promote the degree of frailty of patients,and frailty will react on the treatment of the disease,resulting in the deterioration of the disease.Through the study of frailty,it is found that the relationship between it and existing diseases can assist in the prevention and control of diseases.Therefore,the association between frailty and disease has important research and application value.At present,the related research of frailty is mostly limited to the use of statistical methods of frailty scale to establish data base.In the application process,due to the lack of objectivity,the result error rate is high,while the use of machine learning,data mining and other technologies to model it is less.This article takes cardiovascular disease as an example,and uses unsupervised learning techniques to study the measurement problem of frailty and the relationship between frailty and disease symptoms.The main results achieved in this article are as follows:(1)A measurement model for the degree of frailty of cardiovascular patients based on MGMM is proposed.The model uses KPCA to improve the linear correlation of patient data,and MGMM to improve the accuracy of measuring the degree of frailty of patients.In addition,MGMM is a Gaussian mixture model with multiple iterations.Compared with the measurement method based on the frailty scale,the MGMM-based cardiovascular patient frailty measurement model proposed in this paper has an error rate controlled at about 2.9%,the accuracy rate has increased by about 5.6 percentage points,while the operating efficiency has increased by about 13 percentage points.In addition,the frailty measurement model proposed in this paper provides an accurate data foundation for the study of the relationship between frailty and cardiovascular disease.(2)Based on the proposed MGMM-based frailty degree measurement model,this paper proposes an association rule model of frailty and cardiovascular disease based on the improved Apriori algorithm.Considering that although the basic Apriori algorithm can extract potential association rules between data,it has defects in efficiency and reliability.The model introduces lifting rate and hash table to improve the pruning strategy and data storage structure of Apriori algorithm.The experimental results show that compared with the model using the basic Apriori algorithm,the improved model proposed in this paper improves the operation efficiency by about 60%,and can extract more accurate and effective association rules.To sum up,the model proposed in this paper has better performance in stability,accuracy and efficiency than its comparative model,which proves the feasibility of the model proposed in this paper.At the same time,the model proposed in this paper has a certain degree of scalability,which provides a method reference for the study of the relationship between frailty and other diseases.