摘要
大规模开放在线课程(MOOC)平台作为在线学习的重要工具,能够捕捉丰富的学习行为数据.MOOCube作为关键的数据存储库,汇聚了来自学堂在线的48640个学习行为序列.然而,鉴于MOOC数据的复杂性,传统的序列模式挖掘(SPM)算法在处理这类数据时可能会遇到挑战,导致挖掘出大量不相关的模式.为了提升挖掘结果的有效性,提出了一种效率约束序列模式挖掘(ECSPM)方法.该方法引入了出勤性、离散性和辍学性三大关键约束,这些约束能够精准捕捉学习行为中的不同特征对序列模式挖掘的影响.值得注意的是,这些约束具有向下封闭性质,确保了它们在模式挖掘过程中的有效性.为了发现序列模式(SP),精心设计了3种算法.这些算法结合了逐级搜索空间遍历或递归投影技术,并将成本概念融入模式挖掘过程中.实验评估结果表明:本文提出的算法效果显著.ECSPM不仅显著减少了发现的模式的数量,而且其性能与经典SPM算法相当,甚至在某些方面更优.
Massive Open Online Course(MOOC)platforms have emerged as crucial tools in online learning,capturing vast learning behavior data.MOOCube,a significant data repository,encompasses 48640 learning behavior sequences extracted from XuetangX.However,traditional Sequential Pattern Mining(SPM)algorithms may struggle to handle the complexity of MOOC data,leading to irrelevant patterns.This paper proposes Efficiency-Constrained Sequential Pattern Mining(ECSPM)to facilitate the discovery of useful patterns.Three constraints are introduced:attendance,discreteness,and dropout.These constraints capture the influence of different features of learning behavior on sequential patternsmining.Importantly,these constraints are proven to satisfy the downward closure property,ensuring their effectiveness in shaping the mining process.To discover Sequential Patterns,three algorithms were proposed.These algorithms employ level-bylevel search space traversal or recursive projection techniques while integrating the concept of cost into pattern mining.Experimental evaluation confirms the effectiveness of the proposed algorithm.ECSPM has successfully reduced the number of discovered patterns while maintaining comparable performance to the classical SPM algorithm.
作者
彭亚威
李艳红
杨洋
张法
PENG Yawei;LI Yanhong;YANG Yang;ZHANG Fa(College of Computer Science,South-Central Minzu University,Wuhan 430074,China)
出处
《中南民族大学学报(自然科学版)》
2025年第3期373-383,共11页
Journal of South-Central Minzu University(Natural Science Edition)
基金
湖北省自然科学基金资助项目(2017CFB135)
中央高校基本科研业务费专项资金资助项目(CZY23019)
网络创新及应用型人才课程实践教学研究项目(2019年第一批)。
关键词
学习行为
序列模式挖掘
效率约束序列模式挖掘
成本
learning behavior
sequential pattern mining
efficiency-constrained sequential pattern mining
cost
