Font Size: a A A

Application And Optimization Of Machine Learning In Dynamic Balance Data

Posted on:2021-03-02Degree:MasterType:Thesis
Country:ChinaCandidate:X H LiuFull Text:PDF
GTID:2404330623965005Subject:Computer technology
Abstract/Summary:
Dynamic balance disorder is a common disease in the middle-aged and elderly.The decline of balance ability seriously affects the daily life and mental health of middle-aged and elderly patients.At the same time,due to the decline of the dynamic balance ability of the middle-aged and elderly people,the elderly people are more likely to fall and even die than the young people.Especially in today’s aging world,the social problems caused by the dynamic balance diseases of middle-aged and elderly people are further magnified,which seriously restricts the development of society.Therefore,how to realize the early screening and treatment of middle-aged and elderly people’s sports balance diseases has become a very worthy research topic.In this paper,the method of dynamic balance measurement(evaluation)for middle-aged and elderly people is taken as the research object.Through data mining of more than 17000 collected physical examination data of middle-aged and elderly people’s dynamic balance,two new evaluation models of middle-aged and elderly people’s dynamic balance based on machine learning are constructed,and the two models can be effectively applied to the early screening and treatment of middle-aged and elderly people’s dynamic balance diseases.The research content of this paper mainly includes the following two aspects:Firstly,through the analysis and mining of more than 17000 physical examination data of middle-aged and elderly people’s dynamic balance,a regression evaluation model of middle-aged and elderly people’s dynamic balance is constructed by using RFR-RFE and GBDDTR algorithm.The model only needs to measure the14 indexes of standing on both feet with closed eyes and standing on one foot with closed eyes.It can predict the dynamic balance of middle-aged and old people,and its accuracy meets the clinical requirements.Compared with the traditional method,this method has shorter measurement time and less measurement indicators,which can be well applied to the early screening of human dynamic balance diseases.Secondly,based on the hypothesis of measuring the correlation among actions,48-dimensional measurement indexes in the original data are combined according to actions,and a multi-layer neural network classification model based on LSTM is proposed to evaluate the dynamic balance ability of middle-aged and old people.And through a variety of evaluation indexes,the classification model is evaluated quantitatively.The experimental results verify the validity of the hypothesis and the rationality and validity of the network model.
Keywords/Search Tags:Dynamic Balance, Fall, The Middle-aged and Elderly, Machine Learning, Deep Learning
Related items