基于大数据的思政教学资源推荐系统设计

Design of Recommendation System for Ideological and Political Teaching Resources Based on Big Data

传统的推荐系统只能根据用户的历史数据推荐资源,一旦用户情况发生变化,其推荐精度和推荐效率会快速下降。因此,设计基于大数据的思政教学资源推荐系统。在传统的浏览器/服务器(Browser/Server,B/S)模式上结合Spark计算框架,搭建推荐系统架构。通过分析学习者学习行为、检索行为、资源选用顺序等建立用户对资源类型的偏好画像标签,并以学习者对知识点的掌握情况为依据确定其认知水平;引入热门学习资源选用热度作为惩罚因子,改进协同过滤算法,实现思政资源准确推荐。系统功能测试表明,设计系统响应时间短且推荐资源的准确率高。

The traditional recommendation system can only recommend resources based on the user’s historical data, and once the user’s situation changes, its recommendation accuracy and recommendation efficiency will drop rapidly. Therefore, a recommendation system for civics teaching resources based on big data is designed. The recommendation system architecture is built on the traditional Browser/Server(B/S) model combined with Spark computing framework. By analyzing learners’ learning behaviors, retrieval behaviors,and resource selection order, the user’s preference portrait labels for resource types are established, and their cognitive level is determined based on learners’ mastery of knowledge points;the hotness of popular learning resources selection is introduced as a penalty factor, and the collaborative filtering algorithm is improved to achieve accurate recommendation of civics and political science resources. The system function test shows that the response time of the designed system is short and the accuracy rate of the recommended resources is high.