【作者】 张崇林；

【导师】 虞丽娟；

【作者基本信息】 上海体育学院 ， 运动人体科学， 2012， 博士

【摘要】 研究背景公务员是一类特殊群体，其生活方式、作息制度以及家庭、工作和社会压力等因素使他们长期处在高负荷状态的紧张状态，使他们长期处在亚健康状态，并且是许多慢性病的高发人群。同时，公务员对体育锻炼认同度很高，主观上也有需求，但由于主客观原因，公务员体育锻炼参与度很低，体育锻炼的针对性和科学性差。研究目的本研究以公务员体质健康促进为目的，以人工智能系统的开发为基本理论，以公务员体质健康评价为基础，综合其生活方式、营养状况、作息规律等，研究开发促进公务员体质健康的智能化运动处方系统，为公务员提供智能化、个性化、科学的运动处方，促进其体质健康。研究方法本研究采用文献资料法、专家访谈法、实验测试法、软件工程法、数据挖掘法和系统动力学建模等方法，建立了公务员体质健康促进智能化运动处方系统。结果及结论（1）以人工智能为基本理论，以体质健康评价为基础，开发了促进公务员体质健康的智能化运动处方系统：①提过文献资料和专家访谈，实现了领域的知识获取，并以产生式知识表示法对领域知识进行表示，建立了知识库以及知识推理，并以此建立了规则库；②以体质测试进行数据收集，按推理规则实现数据的转换、存储，建立数据库；③将运动处方作为一个整体，将不同的运动处方作为一个模型，采用框架式知识表示，将运动处方的构成要素用框架结构的“槽”和“槽值”来表示，同时，“槽”和“槽值”也构成了模型的“输入”与“输出”。“槽”和“槽值”之间用产生式推理规则实现其逻辑推理，实现了产生式和框架式知识表示的融合，实现了知识表示和模型库的融合，提高了系统智能化水平；（2）智能化运动处方系统充分考虑公务员的生活方式、作息制度、慢性病等因素，保证了运动处方的可行性、针对性和科学性；（3）研究了数据挖掘理论与方法，着重探讨了基于关联分析数据挖掘和基于人工神经网络数据挖掘理论与技术：①建立了基于体质综合评价人工神经网络模型，为公务员体质综合评价提供了便利算法；建立了基于数据补充的人工神经网络模型，保证了数据库的完整性和有效性；②建立了关联规则数据挖掘模型，为了解不同性别、年龄人群体质特征的提供了方法；数据挖掘对于发现知识、完善知识库和数据库有重要意义，保障了运动处方的效果。（4）以系统动力学理论为基础，以体质为系统，建立了不同性别、年龄人群运动处方的系统动力学模型；通过模拟不同运动处方变量的组合，有针对性的优化了运动处方，提高了运动处方的效果；通过模拟运动周期内体质系统的动态变化，为运动处方执行者提供了实时的数据对照，提高了运动者的运动信心。（5）智能化运动处方系统以人工智能为基础，以Browser/service为背景，充分利用互联网技术，实现了科学化、智能化运动处方的在线服务。更多还原

【Abstract】 BackgroundCivil service is a special group which is under high tension because of their lifestyle, rest and work system and with high pressure coming from jobs and society.Civil service is in sub-health status and is high dangerous group of some chronics.Civil service has a highly recognition about the value of sport for health and hasdesire to participate in it. But because some objective and subjective reasons, civilservice cannot participate in exercise and is lack of targeted and scientific.PurposeThis dissertation has an aim to build an intelligent exercise prescription system inorder to promotion of health of civil service. The system is based theory on artificialintelligence and assessment of fitness-health. The exercise prescription is build by thesystem is intelligent and targeted and scientific.MethodsThis dissertation has some methods to build the intelligent exerciseprescription system of research as well as literature, expert interviews,experimental test, software engineering method, data mining, system dynamicsmodeling.Results and Conclusions(1)The intelligent exercise prescription system based theory on artificialintelligence and assessment of fitness-health include these achievements.①Realzed knowledge acquisition field by literature and expert interviews andrealized the knowledge representation by the method of production knowledgerepresentation in fitness-health. And it established the way of knowledge reasoningand built the knowledge base and rule base.②To collect data by fitness test and realized the switch and memory of datan andbuilt the database.③It is taken the exercise prescription as an entirety and taken different exerciseprescription as one model to realize the knowledge representation of exerciseprescription by frame knowledge representation. It represents the factors of exerciseprescription as “slot” and “slot-value” as while the “slot” and “slot-value” equivalentto the “input” and “output” in model. It realized the logic reasoning in “slot” and“slot-value” by production knowledge representation. So the system realized thefusion knowledge representation between the frame knowledge representation andproduction knowledge representation and realized the fusion between knowledgerepresentation and model base. As a result the system has a much higher level of intelligent.(2) The intelligent exercise prescription system is the life-style and rest/worksystem and chronics of civil service taken into account deeply so as to ensure thefeasibility and targeted and scientific of exercise prescription.(3) On the base of research theory and method of data mining, this dissertationfocus on the application research of the data mining theory and technology based onassociation analysis and artificial neural networks.①To build the artificial neural networks based on fitness comprehensiveevaluation to provide a simple algorithm and build the artificial neural networks basedon filling the lacked data to keep the integrity and effectiveness of database.②To build the model of association rules distinguished by different sexes andages in order to provide a way to learn the characters in different groups. It is veryimportant in knowledge discovery and consummation the database and knowledgebase and ensurance the effectiveness of exercise prescription.(4)On the base of the theory of system dynamics and taken the fitness as asystem this dissertation built some system dynamics models by different sexes andages. The system dynamics models optimize the exercise prescription by simulatingthe varibles combination in exercise prescription and improve the effect. The systemdynamics models provide the real-time data for the agent to urge them participate inexersice in accordance with the exercise prescription by simulating the dynamicchanges of all fitness indexes in the period and which can increase the agent`sconfidance.(5)The intelligent exercise prescription system is dased on the artificialintelligence and sued Browser/service as support and used internet technology torealize the service online of the scientific and intelligent exercise prescriptions.更多还原