基于自组织社区发现的协作学习算法

A Collaborative Learning Algorithm Based on Self-organizing Communities

远程学习者通常很难判断哪些学习资源最适合他们的阅读需要,同时对教师来说,针对每个学习者重新组织不同的学习资源几乎是不可能的。基于此,提出一种新颖的学习偏好建模方法,通过将动态学习数据映射为"资源、评估"的方式实现对学习特征的综合评估;通过构建智能代理来监控学习者的动态学习行为;提出组隶属度奖励机制和组成员交换机制,实现对分布式环境下的相似学习者的社区自组织;同时,基于JADE智能代理平台开发了一个协作学习平台,使得具有相似学习偏好的学习者能够进行学习资源和经验的共享。实验证明,算法具有较高的匹配准确性和社区构建效率,并能够切实提高协作学习的有效性。

E- learners always find it difficult to make a judgement on which learning materials are fit for their need, whilst instructors find it almost impossible to reorganize different materials corresponding to individuals. This paper proposes an innovative learning preference modeling approach which can give the general evaluation on learning features based on the mapping from dynamic learning data to ＂resource, evaluation＂ tuple. Based on intelligent agent, it realizes the monitoring of dynamic learning behaviors . The group membership award and group member exchange mechanism are also proposed in this paper which realize the self - organizing of learner communities in distrib- uted e - learning environment. Furthermore, a collaborative learning platform is implemented based on JADE intelligent agent platform which enables the e - learners to share learning resources and experiences. Experiment results have shown that this algorithm has sustainably improved matching accuracy and community organizing efficiency, and can enhance the validity of collaborative learning.