关键词: DP-DBSCAN算法, 一卡通数据, 关系标签, FP-TRtree模式, 可视化

Abstract: The wide application of campus card technology is an important symbol to measure the informatization degree in colleges. Among them, the students’ consumption data imply a great potential value, and it is a significance to excavate. Motivated by this problem, this paper proposes a method to convert campus streaming data into a consumption trajectory tree DP-DBSCAN algorithm with spatial attribute, and builds FP-TRtree mining model with relationship label. DP-DBSCAN algorithm adopts the time block, order query and distance measurement, it can transfer data into FP-TRtree frequent item set with order effectively, ignoring parameters. The FP-TRtree model adds value in sequence, supports descending order, and optimizes the relationship label between the same loci. Visualization analysis demonstrates that the method and model not only find the student relationship track network of frequent consumption and isolated populations, but also make a quantitative description between nodes of students spending intimate degree. At the same time, the method reduces the database scan times and the establishment of the branch of a tree. The experimental results conform to real consumption of students, it can find hidden information from the complex consumption network and provide the basis for school management.

Key words: DP-DBSCAN algorithm, cartoon data, relationship label, FP-TRtree mode, visualization