MOOC学习行为数据中因果关系的挖掘方法
|
摘要
目前,MOOC数据挖掘关注重点是学习效果和学习行为的相关性。但是相关性并不意味着因果关系。后者更有助于为构建智能化导学、推荐、评价等机制提供依据。为此,提出了一种"因果图构建+隐变量插入"相结合的因果关系挖掘方法。该方法首先按"无向图学习+方向学习"两阶段,从高维海量的学习行为数据中构建出因果图,然后生成隐变量候选集,并基于因果图中的半团结构将隐变量插入因果图,从而获得简化和易理解的因果关系。
At present, the focus of MOOC data mining is the correlation between learning effect and learning behavior. However, the correlation doesn't mean causality. The latter will provide bases more helpfully for those like constructing intelligent guidance, recommendation, evaluation and so on. Therefore, this paper proposes a causation mining method, which combines causality graph construction with implicit variable insertion. This approach first constructs a causality graph based on the high dimensional and massive data on learning behavior according to the undirected graph learning and directional graph learning. Then, the candidate set of latent variables are generated, and the latent variables are inserted into the causal graph based on the half-cluster structure in the causality graph to obtain a simplified and easy-to-understand causal relationship.
At present, the focus of MOOC data mining is the correlation between learning effect and learning behavior. However, the correlation doesn't mean causality. The latter will provide bases more helpfully for those like constructing intelligent guidance, recommendation, evaluation and so on. Therefore, this paper proposes a causation mining method, which combines causality graph construction with implicit variable insertion. This approach first constructs a causality graph based on the high dimensional and massive data on learning behavior according to the undirected graph learning and directional graph learning. Then, the candidate set of latent variables are generated, and the latent variables are inserted into the causal graph based on the half-cluster structure in the causality graph to obtain a simplified and easy-to-understand causal relationship.
引文
[1]Shapiro H,Lee C H,Roth N E,et al.Understanding the Massive Open Online Courses(MOOC)Student Experience:An Examination of Attitudes,Motivations,and Barriers[J].Computers&Education,2017(110):35-50.
[2]Alshabandar R,Hussain A J,Laws A,et al.Machine Learning Approaches to Predict Learning Outcomes in Massive Open Online Courses[C]//International Symposium on Neural Networks IEEE,2017:713-720.
[3]蒋卓轩,张岩,李晓明.基于MOOC数据的学习行为分析与预测[J].计算机研究与发展,2015,52(3):614-628.
[4]Sugihara G,May R,Ye H,et al.Detecting Causality in Complex Ecosystems[J].Science,2012,338(6106):496-500.
[5]Justin R.Rebooting MOOC Research[J].Science,2015,347(6217):34-35.
[6]付剑锋,刘宗田,刘炜,等.基于层叠条件随机场的事件因果关系抽取[J].模式识别与人工智能,2011,24(4):567-573.
[7]Thadhani R,Appelbaum E,Pritchett Y,et al.Vitamin DTherapy and Cardiac Structure and Function in Patients with Chronic Kidney Disease:the PRIMO Randomized Controlled Trial[J].Jama,2012,307(7):674-684.
[8]Worm B S,Jensen K.Does Peer Learning or Higher Levels of e-learning Improve Learning Abilities?a Randomized Controlled Trial[J].Medical Education Online,2013,18(1):21877.
[9]Lager A C J,Torssander J.Causal Effect of Education on Mortality in a Quasi-experiment on 1.2 Million Swedes[J].Proceedings of the National Academy of Sciences,2012,109(22):8461-8466.
[10]蔡瑞初,陈薇,张坤,等.基于非时序观察数据的因果关系发现综述.计算机学报,2017,40(6):1470-1490.
[11]Hyttinen A,Eberhardt F,Hoyer P O.Learning Linear Cyclic Causal Models with Latent Variables[J].Journal of Machine Learning Research,2012,13:3387-3439.
[12]Colombo D,Maathuis M H,Kalisch M,et al.Learning High-dimensional Directed Acyclic Graphs with Latent and Selection Variables[J].The Annals of Statistics,2012,40(1):294-321.
[13]王晓宇,欧阳丹彤,赵剑.不完备模型下的离散事件系统诊断方法[J].软件学报,2012,23(3):465-475.
[2]Alshabandar R,Hussain A J,Laws A,et al.Machine Learning Approaches to Predict Learning Outcomes in Massive Open Online Courses[C]//International Symposium on Neural Networks IEEE,2017:713-720.
[3]蒋卓轩,张岩,李晓明.基于MOOC数据的学习行为分析与预测[J].计算机研究与发展,2015,52(3):614-628.
[4]Sugihara G,May R,Ye H,et al.Detecting Causality in Complex Ecosystems[J].Science,2012,338(6106):496-500.
[5]Justin R.Rebooting MOOC Research[J].Science,2015,347(6217):34-35.
[6]付剑锋,刘宗田,刘炜,等.基于层叠条件随机场的事件因果关系抽取[J].模式识别与人工智能,2011,24(4):567-573.
[7]Thadhani R,Appelbaum E,Pritchett Y,et al.Vitamin DTherapy and Cardiac Structure and Function in Patients with Chronic Kidney Disease:the PRIMO Randomized Controlled Trial[J].Jama,2012,307(7):674-684.
[8]Worm B S,Jensen K.Does Peer Learning or Higher Levels of e-learning Improve Learning Abilities?a Randomized Controlled Trial[J].Medical Education Online,2013,18(1):21877.
[9]Lager A C J,Torssander J.Causal Effect of Education on Mortality in a Quasi-experiment on 1.2 Million Swedes[J].Proceedings of the National Academy of Sciences,2012,109(22):8461-8466.
[10]蔡瑞初,陈薇,张坤,等.基于非时序观察数据的因果关系发现综述.计算机学报,2017,40(6):1470-1490.
[11]Hyttinen A,Eberhardt F,Hoyer P O.Learning Linear Cyclic Causal Models with Latent Variables[J].Journal of Machine Learning Research,2012,13:3387-3439.
[12]Colombo D,Maathuis M H,Kalisch M,et al.Learning High-dimensional Directed Acyclic Graphs with Latent and Selection Variables[J].The Annals of Statistics,2012,40(1):294-321.
[13]王晓宇,欧阳丹彤,赵剑.不完备模型下的离散事件系统诊断方法[J].软件学报,2012,23(3):465-475.