基于Consume2Vec模型的校园一卡通大数据分析  

Consume2Vec model-based analysis of campus card big data

作　　者：韩泽峰 杨涛 侯琳琳 田强[4] 刘良金 吴偶 HAN Zefeng;YANG Tao;HOU Linlin;TIAN Qiang;LIU Liangjin;WU Ou(Center of Applied Mathematics,Tianjin University,Tianjin 300072,China;Research Center of Big Educational Data,Hangzhou Zhihuzheye Company,Hangzhou Zhejiang 310008,China;Center for Combinatorics,Nankai University,Tianjin 300072,China;Administration Center,Tianjin Normal University,Tianjin 300387,China)

机构地区：[1]天津大学应用数学中心,天津300072 [2]杭州知乎者也科技有限公司教育大数据研发中心,杭州310008 [3]南开大学组合数学中心,天津300072 [4]天津师范大学行政管理中心,天津300387

出　　处：《计算机应用》2020年第S01期85-91,共7页journal of Computer Applications

基　　金：国家自然科学基金资助项目(61673377);天津人工智能专项(17ZXRGGX00150)。

摘　　要：现有的一卡通数据挖掘方法大多采用统计、聚类、关联规则等浅层机器学习和数据挖掘方法,忽略了消费数据的时序性,缺乏对数据的深度表达。基于深度神经网络,提出了能够对消费数据时序性和关联性进行深度挖掘的Consume2Vec模型,并在此模型的基础上构建消费异常检测模型。通过在大规模一卡通消费数据上进行实验,验证了两个具体Consume2Vec模型的性能,并从不同维度将学生划分为不同群体进行对比分析,发现学生的消费规律和特点。Most of the existing campus card data mining methods use shallow machine learning and data mining methods such as statistics,clustering,and association rules,ignoring the timing of consumption data and lacking deep expression of data. This paper proposed a Consume2Vec model based on deep neural network,which can deeply mine the timing and correlation among consumption data,and build a consumption anomaly detection model based on Consume2Vec model.Experiments were conducted on a large-scale card consumption data,verify the performance of two specific Consume2Vec models,and divide the students into different groups from different dimensions for comparative analysis,and find some consumption rules and characteristics of students.

关 键 词：校园一卡通大数据 TRANSFORMER 长短期记忆 局部异常因子算法 消费异常检测 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]