Analysis Of The Willingness Of The Elderly To Live In Institutions Based On Machine Learnin

Population aging is a stage that society will inevitably experience to a certain extent.By the end of 2015,the proportion of senior citizens aged 60 and above had reached 16.15%,and by the end of 2021,that number had risen to 18.9%.The proportion of the elderly population has risen rapidly in a few years.It can be seen that our country not only enters a serious Aging society but the process of aging is fast.With the acceleration of aging society,the pension problem facing our country is increasingly thorny,and due to the development of economy and the continuous improvement of living standard,people’s requirements for the pension have also improved.“ Getting old before getting rich ” has brought political,economic and social pressures to our country.Therefore,there is a mitigation strategy to build a pension service system based on home care,community care and institutional care.This study is conducive to changing the traditional home-based care,liberating the young and middle-aged labor and providing better elderly care services,policy formulation and accurate publicity for nursing homes.The data used in this article is from the results of The Chinese Longitudinal Healthy Longevity Survey(CLHLS)in 2018.The questionnaire involved more than500 sample sites in 22 provinces.The choice of whether or not to go to pension institutions in the results was selected as the dependent variable,and the other 17 questions related to the research content were selected as the independent variable.The data is then preprocessed,including missing value processing,SMOTE oversampling processing,one-hot coding,and normalization.Models are constructed using logistic regression,support vector machines,and random forest algorithms in machine learning.Models are trained by Python and the Sklearn package,and we optimize the parameters of the model.We evaluate and compare models with precision,recall,F1-score,accuracy,and AUC.We analyze and predict data with the hope that the results can provide data support for future related studies.It can be seen through the evaluation of models that the precision of all three models can reach more than 90%.with the compositive consideration of evaluation of models,the results of random forest are the best.The results of support vector machine are second and it’s been running for a long time.The precision of logistic regression is the lowest among the three models,but its model principle is the simplest.The logistic regression butterfly diagram intuitively shows the characteristics that have a great influence on the willingness of the elderly to live in pension institutions.Among them,the people with the strongest tendency to live in pension institutions are the elderly who now live with strangers.Some people who are in a state of divorce or have enough living expenses tend to live in pension institutions.It can be seen that the elderly with good economic conditions or live inconveniently on their own are more inclined to go to pension institutions.The willingness of the elderly from different backgrounds for institutional care varies greatly.