基于本体兴趣特征向量空间模型的社区自组织算法  

A Self-organization Algorithm for Community Based on Ontology-based Learner's Interest Eigenvector Space Model

作　　者：程艳[1] 许维胜[1] 杨继君[2] 何一文[1] 

机构地区：[1]同济大学电子与信息工程学院,上海201804 [2]同济大学经济与管理学院,上海201804

出　　处：《系统工程》2009年第5期96-103,共8页Systems Engineering

基　　金：国家自然科学基金资助项目(70871091;60804042)

摘　　要：为了解决远程教育不可避免地产生的"孤独"学习者的问题,把具有相同学习兴趣的学习者组织到同一个学习社区中进行协作式学习。学习社区建立的重点和难点在于学习者之间相似关系的判定和计算,针对传统的向量空间模型中术语间语义相关性被忽略的不足,提出基于本体的向量空间模型来计算学习者的兴趣特征向量,根据兴趣的隐性表示获取对应的显式表示,此计算模型提高了兴趣相似性比较的精确程度。同时提出了一种基于学习者兴趣相似匹配度和学习者兴趣匹配浓度的学习社区的自组织算法。针对基于本体的向量空间模型使用本体中的概念构造向量空间表现出的巨大维数,运用概念索引降维法对兴趣特征矩阵进行合理降维,大大减少了计算的复杂性。最后,以网络学习案例来进行实验分析,验证该模型算法具有较高的效率和良好的扩展性。To settle the unavoidable problem of loneliness from the long-distance education and to help the students learn more cooperatively, learners of the same interest can be organized into one study community. The key to establishing a learning community is to determine and calculate the similarity between the learners. To get rid of the disadvantages of neglecting the semantic relevance between terms in the traditional vector space model, ontology-based vector space model is presented to calculate the learner＇s interest eigenvector, which can acquire the corresponding explicit express （that is, vectors of interest） according to the recessive expression and enhance the relative accuracy of the interest similarity. And a self-organization algorithm is put forward, based on the learners＇ interest similarity match-degree and its concentration. Great dimensions would take place with the ontology to construct vector space, thus Concept Indexing method and reasonable treatment to matrix of interest Eigen value are used here to promote the calculation efficiency. Finally, an experimental analysis of online education cases is carried out to verify the model algorithm with high efficiency and excellent expansibility.

关 键 词：本体 兴趣特征向量空间模型 概念索引降维 兴趣相似匹配度 兴趣匹配浓度 

分 类 号：G424[文化科学—课程与教学论]