教育信息化2.0时代Web学习资源多特征融合的联合排序推荐模型研究

教育信息化1.0时代对Web网页学习资源的搜索排序与推荐，存在着对不同网络学习者千篇一律呈现相同或相似资源的不足；而教育信息化2.0或3.0时代是一个“互联网+教育“的技术驱动时代，其中资源排序推荐技术需实现Web学习资源与网络学习者之间“排序推荐”的精准化与个性化，这在花冠病毒肆虐并实施在线教学的当下以及后疫情时代，其理论研究意义与实践价值明显。因此本课题提出了“教育信息化2.0时代Web学习资源多特征融合的联合排序推荐模型”研究问题。在基于传统网页内容排序、链接结构排序、用户反馈特征排序和网页可信度排序等四类独立的基础网页排序算法基础上，原创设计并开发自有产权的网络爬虫系统抓取Web网页大数据，创建四大类传统排序算法的各自排序索引库，然后依据马氏距、高斯核函数和Felder-Silverman学习风格量表为基础建立全连接网络结构图，应用半监督自动学习机原理动态拟合多特征融合的Web学习资源联合排序推荐的动态交互过程并进行建模与反馈评估。

In the era of educational informatization 1.0, there is a shortage of identical or similar resources for different web learners. In the era of educational informatization 2.0 or 3.0 driven by educational intelligent technology, in which Web resource ranking recommendation technology needs to achieve the accurate and individualized of requirements between Web learning resources and its learners. Therefore the research problem of "Joint Ranking and Recommendation Model of Multi-feature Fusion of Web Learning Resources in the Era of Educational informatization 2.0" is put forward. On the basis of four independent basic web page sorting algorithms relatively, i.e. traditional web page content sorting, link structure sorting, user feedback feature sorting and web page reliability sorting, an independent web crawler tool is designed and developed to capture web page data, respective web page sorting index libraries of traditional sorting algorithm's are created. Then, according to Markov distance, Gauss kernel function and Felder-Silverman, the sorting index libraries of four kinds of traditional sorting algorithms are established and used to construct a fully connected network structure chart. The principle of distributed semi-supervised automated learning machine is applied to dynamically fit the recommendation process of Web learning resources ranking based on multi-feature fusion. Finally, the reliability of the joint ranking recommendation results data is evaluated by using big data processing and analysis tools.