【作者】 陈仕品；

【导师】 张剑平；

【作者基本信息】 西南大学 ， 教育技术学， 2009， 博士

【摘要】 适应性学习支持系统是近年来人工智能在教育中应用的研究热点之一,它是教育学、认知科学和计算机科学的交叉研究领域。由于传统网络教学系统忽视了学习者的个体差异,对所有的学习者采用相同的教学内容和教学策略,导致了学习者经常发生网络迷航和认知超载等问题,在很大程度上影响了网络教学的质量,因此目前许多研究开始关注创建适应性和个性化的学习环境。适应性学习支持系统在本质上是一类支持个性化学习的在线学习环境,它能够针对个体在学习过程中的差异性而提供适合个体特征的学习支持,包括个性化的学习资源、学习过程和学习策略等。本研究针对适应性学习支持系统的学生模型和适应性教学内容的组织开展研究。在适应性学习支持系统中,学生模型是系统的核心组件,它记录学习者的个体特征,反映了学习者的个体差异,为系统进行智能决策提供了决策依据;适应性教学内容是系统根据学习者个体特征定制的个性化教学内容,它既是系统最重要的学习支持方式,也是系统适应性最直接的体现。本研究采用了文献研究法、系统方法和基于设计的研究方法。研究工作主要包括:1.以历史发展为线索总结了计算机辅助教学的智能化历程,分析了智能教学系统的优势与不足,认为适应性学习支持系统是当前数字化学习支持平台的发展趋势。2.在深入分析增强适应性超媒体应用模型(EAHAM)的基础上,提出了一种基于EAHAM模型的适应性学习支持系统的体系结构,它主要包括媒体空间、领域知识模型、学生模型、情境模型、教学模型和适应性模型等六个部分,并添加了适应性学习模块、学习策略模块和学习工具模块。基于EAHAM模型的适应性学习支持系统的体系结构深化了系统的组成部件和运行机制,不仅能够根据学习者在知识基础、学习风格等方面的个体差异提供适应性学习支持,而且在系统实现方面具有良好的可操作性。3.建立了一种基于认知状态和学习风格的学生模型,反映了学习者在认知状态和学习风格两方面的个体差异。通过对智能教学系统中典型的学生模型进行了详细分析,认为智能教学系统中的学生模型主要功能是诊断并记录学生的知识状态,特别是诊断学生在问题解决过程中形成的错误概念。这种学生模型局限于学习者的知识状态,而对学习者的其它个体特征缺乏了解。为了反映学习者在先前知识基础和学习风格方面的个体差异,本研究提出了一种新的基于认知状态和学习风格的学生模型,它主要包括了学生描述、学习风格、认知状态和学习历史。其中,学习者的认知状态和学习风格是基于EAHAM模型的适应性学习支持系统的主要适应维度。本研究采用Felder-Silverman学习风格模型(FSLSM),通过学习者在网络课程在线注册时填写问卷调查进行初始化。4.将认知能力和学习风格作为基于EAHAM模型的适应性学习支持系统的两个适应性维度,提出了基于认知能力和学习风格的适应性教学内容组织模型,详细分析了适应性教学内容的组织过程。适应性教学内容的动态组织模型主要包括两个过程:一是适应性学习支持系统根据领域知识的层级结构、面向任务的教学策略和学习者的认知能力水平分类动态地组织适应性教学内容。二是适应性学习支持系统根据FSLSM的“感知——输入”两个维度提供适应性教学内容表示策略:根据FSLSM的“处理——理解”两个维度提供适应性导航策略。适应性教学内容采用SCORM标准进行封装,并在学习过程中基于认知状态进行适应性标注。5.采用计算机自适应测试技术来诊断学习者的认知能力,并根据测试结果动态地更新学生模型。在项目反映理论基础上,提出了一种自适应在线测试系统的体系结构,采用三参数逻辑斯蒂模型分析了选题算法、能力评估算法和测试终止条件,并针对《现代教育技术》国家精品课程设计了自适应测试的原型系统MET-CATS,分析了系统自适应测试的运行过程和评价过程。根据测试结果,学习者的认知能力被分为初级水平、中级水平和高级水平。更多还原

【Abstract】 Adaptive learning support system has been the focus topic in the research field of artificial intelligence in education in recent years, which is a cross areas of education, cognitive science and computer science. However, current e-learning system sends the same teaching content to all learners without concerning individual differences of learners, such as the prior knowledge, learning goal and learning style, which results in the phenomenon of cognitive overload and disorientation, and seriously affects the quality of e-learning. So many researches shift from tranditional e-learning platform to adaptive and personalized learning environment. Adaptive learning support system in essence is a kind of personalized online learning environment, which can provide adaptive learning support according to individual characteristics of learner, including personalized learning resources, learning processes and learning strategies.In this paper, we focus on the student model and adaptive teaching content organization of adaptive learning support system. Student model is the key component of adaptive learning support system, which records the learner’s individual characteristics, reflecting the individual differences of learners, and provides a basis for decision making of adaptive learning support system. And adaptive teaching content is customized in accordance with the individual characteristics of learners by adaptive learning support system, which is the most important way to support learning, and is the most obvious manifestation of adaptation of adaptive learning system.In this paper, the literature, systematic approach and design-based research methods are adopted. The main works of this research as follows:Firstly, we summarize the being intelligent process of computer-aided instruction based on literature review, and analyze the strengths and weaknesses of intelligent tutoring system, and hold that adaptive learning support system is the current trend of the e-learning platform.Secondly, based on detailed analysis of the Enhancing Adaptive Hypermedia Application Model (EAHAM), we proposes the architecture of adaptive learning support system, including the media space, domain knowledge model, student model, context model, teaching strategies model and adaptive model, and adding the adaptive learning module, learning strategies modules and learning tools modules. The architecture of adaptive learning support system based on EAHAM refines the system components and the operating mechanisms, which can provide adaptive learning support in accordance with the prior knowledge, learning style and other aspects of individual differences of learners, and has good operability in system implementation.Thirdly, we set up the student model based on cognitive state and learning style, which records the individual differences of learners. Through the detailed analysis of typical student model in intelligent tutoring system, we consider that the main functions of student model in intelligent tutoring system is diagnosing and recording the student’s knowledge state, especially the misconceptions in problem-solving process. This student model is confined to the learner’s knowledge state, whereas it lacks of understanding other individual characteristics of learners. In order to reflect the learner’s individual differences, such as the prior knowledge, learning goal, and learning style, we put forward a student model based on cognitive state and learning style. New student model mainly includes the student profiles, learning style, cognitive state, as well as learning history. Among them, learner’s cognitive state and learning style is the major adaptive dimensions of EAHAM-based adaptive learning support system. In this paper, in order to describe the learning style of learner, Felder-Silverman learning style model (FSLSM) is adopted. When learner registers in adaptive learning support system, learning style is initialized through learner filling in the ILS questionnaire.Fourthly, according to the individual differences of cognitive state and learning style in student model, this paper puts forward the dynamically generation process model of adaptive teaching content, which includes both the process. Firstly, system dynamically generates teaching content sequences in accordance with the cognitive abilities of learners. Based on the hierarchical structure of domain knowledge, the task- centered teaching strategies and classification of learners’ cognitive ability level, adaptive teaching content is dynamically organized. Secondly, Adaptive learning support system provides adaptive presentation strategies of teaching content in accordance with the "perception - input" dimensions of FSLSM, and provides adaptive navigation strategies in accordance with the "processing - understanding" dimensions of FSLSM. Adaptive teaching content is packaged by SCORM standards, and is adaptively annotated based on learner’s cognitive states in the learning process.Fifthly, we diagnose the learner’s cognitive ability based on computerized adaptive testing technology, and dynamically update student model in accordance with the test results. Based on three-parameters logistic model, the architecture of adaptive online testing system is proposed, and item bank, item selection algorithm, ability evaluation algorithm and test termination conditions are analyzed. As well as against National Excellent Courses "modern educational technology", we develop the MET-CATS prototype of computerized adaptive testing system, and analyze its evaluation process. According to the results of computerized adaptive testing, system will be divided cognitive abilities of learners into the primary level, intermediate level and advanced level.更多还原