基于BP神经网络对北京市社区中医药服务发展影响因素研究

发布时间：2018-04-17 08:11

本文选题：BP神经网络 + 服务发展　；参考：《北京中医药大学》2017年硕士论文

【摘要】：目的:通过文献梳理北京市社区中医药服务发展中的问题,结合社区中医药相关政策,多角度构建社区中医药服务发展影响因素初步框架。运用BP神经网络方法建立社区中医药服务发展影响因素分析模型,计算北京市社区中医药服务发展影响因素指标权重,分析北京市社区中医药服务发展中的主要影响因素,探讨该方法在影响因素分析中的优势。方法:1、文献研究法。收集CNKI,万方,维普等数据库中基层中医药文献,归纳2005年至2016年社区中医药服务发展现状及存在的问题。2、问卷调查法。以函调和现场调研方式,收集北京市64家社区卫生服务中心中医药服务数据。3、BP神经网络分析。本研究采用BP神经网络方法分析北京市社区中医药服务发展影响因素权重。利用matlab 2010b软件,构架以北京市社区中医药服务发展影响因素为输入变量,社区中医药服务发展效率为输出变量的三层网络模型,通过相关公式转化成社区中医药服务发展影响因素权重值。4、DEA分析。运用超效率CCR模型,测算社区中医药服务发展效率,将社区中医药发展效率作为BP神经网络模型输出变量。5、频数统计分析。对基本情况、人员、服务中部分指标和制约因素采取频数统计方法进行分析,为本研究建议的提出提供数据支持。结果:本研究建立了以社区中医药服务发展影响因素指标为输入节点,以社区中医药服务发展效率为输出节点,隐节点数为8个的三层神经网络模型。其中,样本数为64例,权重结果如下:中医药业务用房面积(0.0974),中药饮片种类(0.1002),中医药设备种类(0.1118),中医师数(0.1006),中级以上中医师数(0.1376),中医药适宜技术种类(0.1250),新开展的中医药适宜技术种类(0.1174),重点人群中医药保健种类(0.0964),中医药慢病管理种类(0.1136)。结论:1、人员、技术是北京市社区中医药服务发展主要影响因素。北京市社区中医药影响因素权重值前四位的因素为中级以上中医师数、中医药适宜技术种类、新开展的中医药适宜技术种类、中医药慢病管理种类。2、BP网络是一种优质的分析社区中医药服务发展影响因素的方法。在运用两种方法构建社区中医药服务发展影响因素模型时,BP神经网络模型R2值达到0.97左右,而多元线性回归模型R2仅为0.1991,回归模型各自变量P值大于0.05,模型没有统计学意义。
[Abstract]:Objective: to analyze the problems in the development of community traditional Chinese medicine (TCM) services in Beijing, and to construct a preliminary framework of influencing factors for the development of community traditional Chinese medicine (TCM) from different angles.Using BP neural network method to establish the analysis model of influencing factors of community traditional Chinese medicine service development, to calculate the index weight of influencing factors of community traditional Chinese medicine service development in Beijing, and to analyze the main influencing factors in the development of community traditional Chinese medicine service in Beijing.The advantages of this method in the analysis of influencing factors are discussed.Methods: 1, literature research.The basic TCM documents in CNKI, Wanfang and Weipu databases were collected, and the status quo and existing problems of community TCM service development from 2005 to 2016 were summarized.The data of traditional Chinese medicine service in 64 community health service centers in Beijing were collected by correspondence and field investigation. BP neural network was used to analyze the data.In this study, BP neural network method was used to analyze the weight of factors influencing the development of community traditional Chinese medicine service in Beijing.By using matlab 2010b software, a three-layer network model with the factors influencing the development of community traditional Chinese medicine service in Beijing as input variable and the efficiency of community traditional Chinese medicine service development as output variable is constructed.The weight value of influencing factors of community traditional Chinese medicine service development was transformed into DEA analysis by relevant formulas.The development efficiency of community traditional Chinese medicine (TCM) service is calculated by using the super-efficiency CCR model. The development efficiency of community TCM is regarded as the output variable of BP neural network model .5. the frequency is statistically analyzed.The basic situation, personnel, some indicators and constraints in the service are analyzed by means of frequency statistics to provide data support for the proposal of this study.Results: in this study, a three-layer neural network model was established with the index of influencing factors of community TCM service development as the input node, the community TCM service development efficiency as the output node and the number of hidden nodes as 8 nodes.Among them, the sample size is 64,The weight results are as follows: the area of accommodation used in Chinese medicine business is 0.0974m, the type of Chinese medicine pieces is 0.1002U, the type of equipment of Chinese medicine is 0.1118m, the number of TCM doctors is 0.1006m, the number of doctors above intermediate level is 0.1376m, the category of suitable technology of traditional Chinese medicine is 0.1250m, the newly developed type of suitable technology of traditional Chinese medicine is 0.1174.The type of health care of Chinese medicine is 0.0964, and the type of management of chronic disease of traditional Chinese medicine is 0.1136.Conclusion 1, personnel and technology are the main influencing factors for the development of community traditional Chinese medicine service in Beijing.The first four factors of the weight value of the influencing factors of traditional Chinese medicine in Beijing community are the number of Chinese medicine doctors at or above the intermediate level, the types of appropriate techniques of traditional Chinese medicine, and the new types of suitable techniques of traditional Chinese medicine.BP network is a good method to analyze the influencing factors of community TCM service development.When two methods were used to construct the model of influencing factors of community TCM service development, the R2 value of BP neural network model was about 0.97, while that of multivariate linear regression model was only 0.1991.The regression model's variables P value was more than 0.05, and the model had no statistical significance.
【学位授予单位】：北京中医药大学
【学位级别】：硕士
【学位授予年份】：2017
【分类号】：R197.61

【参考文献】