面向虚拟学习社区的兴趣驱动个性化学习新模式研究-以江西教育为例

项目名称： 面向虚拟学习社区的兴趣驱动个性化学习新模式研究-以江西教育为例

项目编号： No.61262080

项目类型： 地区科学基金项目

立项/批准年度： 2013

项目学科： 自动化技术、计算机技术

项目作者： 程艳

作者单位： 江西师范大学

项目金额： 43万元

中文摘要： 当前我国"就业难"和"人才荒"两难并存的现象尤为突出，个性化人才的培养迫在眉睫。面向虚拟学习社区，有机结合心理测量学、认知诊断、系统动力学、控制理论、人工智能和计算机科学等多学科知识，探寻高效的个性化新模式，以求理论与方法上的新突破，探究教与学的深层规律。借用行为动力学方法，建立兴趣驱动的行为动力学模型，揭示兴趣和行为相互作用机制，进行潜在兴趣预测；基于项目反应理论(IRT)与DINA认知诊断模型，建立兴趣与能力两者之间关联关系模型，厘清兴趣与能力之间的关系，建构定量模型；基于agent建模和行为控制方法建立兴趣驱动行为控制机制，构建行为网络模型，探寻高效行为选择算法。通过研究个体和群体行为特征，描述行为网络体系的复杂行为特征及逻辑关系，探寻个体与群体高效交互机制，揭示个体之间自组织产生的系统整体兴趣演化规律。面向虚拟学习社区的学习行为动力学分析也为人类行为非泊松特性提供实证研究案例。

中文关键词： 人类行为动力学；幂律；认知诊断；兴趣驱动；个性化学习

英文摘要： Currently, great employment pressure and lack of talents are especially prominent phenomenon in China, so it is urgent to cultivate individualized talents. Combining psychometrics, cognitive diagnosis, system dynamics, controlling theory, artificial intelligence and computer science organically, new efficient individualized patterns and deep laws between teaching and learning in virtual learning community are explored so as to make breakthroughs in theory and methods. Adopting methods of behavior dynamics, a behavior dynamic model driven by interests is built to reveal the interaction mechanism between interest and behavior and to predict potential interest. The model, on the basis of Item Response Theory(IRT)and DINA model in cognitive diagnosis, shows the relationship and difference between interest and ability, thus a quantitative model. A behavior network model, based on agent and behavior control, and behavior control mechanism driven by interest is also adopted to explore an efficient calculation method. Through individual and group behavior characteristics, and the description of online behavior system and logic, we are seeking for an efficient interaction mechanism to make out the deducing laws of interest among individuals in self-organized system. The analysis of learning behavior dynamics in virtual

英文关键词： human dynamics；power-law；cognitive diagnosis；driven by interests；personalized learning