基于Hebbian法则的P2P学习网络构建算法  

An Hebbian Learning Based Algorithm for Constructing P2P Learning Network

作　　者：陈华曦[1] 金晶[1] 陈清华[1] 

机构地区：[1]上海交通大学计算科学与工程系,上海200240

出　　处：《计算机仿真》2009年第2期315-319,361,共6页Computer Simulation

基　　金：国家自然科学基金项目(60372078)

摘　　要：针对远程教育环境中信息量过大,学习者分散、难于组织的问题,提出了一种E-Learning学习网络构建算法。该算法采用P2P的网络结构组织节点之间的通信,利用Hebbian学习法则来修改网络节点之间的信任权值,通过不断调整节点的联系人来实现学习网络的构建。学习网络能够对学习者关于学习资源的查询做出符合学习者学习兴趣的反馈,从而最大限度地满足学习者的需求,提高学习者的学习满意度。实验结果表明,算法相比于传统算法,具有较快的网络建设速度、较高的网络建设质量以及较高的鲁棒性,因而能够更好地适用于大规模的E-Learning学习这种开放的分布式环境。In this paper, an algorithm to construct E -Learning network is proposed to address the problem of information overload and the difficulty of organizing distributed learners in E - Learning environment. The algorithm uses P2P structure to organize communications among nodes, adopts Hebbian learning law to adjust the trust weight between network nodes and finally constructs the network by constant adjustment of contracts of the nodes. Eventually, learning network will be able to feed back appropriate information to a learner＇s query according to its learning interest, thus satisfying the needs of learners to the maximum extent. Experiments show that, compared with traditional algorithm, algorithm in this paper achieves higher construction speed, a better construction quality and a better robustness, and is more suitable for the large - scale E - Learning environment, which is open and distributed.

关 键 词：海布学习法则 远程学习 点对点 学习网络 

分 类 号：TP391.7[自动化与计算机技术—计算机应用技术]