教育大数据是近年来的研究热点。利用分布式文件系统对校园一卡通数据进行存储、预处理和分布式计算。在此基础上,提出学生生活轨迹中的相遇模型,从而挖掘学生线下社交关系。为了区分熟悉的陌生人和真正的好友,从单个学生和整个班级两个角度分析学生线下相遇行为,既可以挖掘好友关系(包括比较孤立的学生),又可以为校园班级社群管理提供数据支撑。实验结果表明,挖掘出的社交关系比较符合实际情况。
Recently, educational big data has become a hot topic. A distributed file system to store, preprocess, and analyze campus card data was adopted. Based on it, a student encounter model has been proposed, so as to mine students" offline social relations. To distinguish real friends from familiar strangers, the offline social relations for either individual students or classes were analyzed. Through these two perspectives, the students' offline encounters was analyzed, which can not only extract social relationship between friends (including the isolated students), but also provide data support for the campus class management. The experimental results show that the mined social relations reflect the real relationship.