摘要： 在网络环境下，Web教育资源规模日益庞大，用户申请资源的过程逐渐复杂化。为此，提出一种基于Agent的Web教育资源预选择分层模型。根据预选择分层模型对Web教育资源进行两层过滤，利用基于语义相似度的过滤算法，将Web教育资源根据语义相似度完成匹配筛选；采用用户反馈信息建立机器学习模型，使用基于Q学习的过滤算法筛选Web教育资源。实验结果表明，分层模型可供用户选取符合用户需求的资源，具有较好的可扩展性。

关键词: Web教育资源, 资源过滤, 语义相似度, Agent技术, 机器学习, 分层模型

Abstract: Towards more and more complexity of choosing the expected resource from large-scale resources in the Web environment, an Agent-based pre-selection hierarchical model of Web education resource is presented. The pre-selection hierarchical model is a two-layer filtering model: The filtering algorithm based on semantic similarity is used to match and filter the Web education resources with semantic similarity; The filtering algorithm based on Q learning is used to establish a machine learning model for fine filtration of Web education resources according to users’ feedbacks. Experimental results show that this model can select the resources meeting users’ demands and it has excellent scalability.

Key words: Web education resource, resource filtering, semantic similarity, Agent technology, machine learning, hierarchical model

中图分类号: 

• N945.12