Using Machine Learning to Understand Students' Learning Patterns in Simulations

This study explores how unsupervised machine learning (ML) techniques can identify productive learning patterns as students conduct virtual experiments using a simulation. The log data from 52 pairs of eighth-grade students were analyzed using two ML techniques, Finite Mixture Model (FMM) and Sequential Pattern Mining (SPM). The results show four levels of learning patterns (i.e., Low Activity, Random Interaction, High Analysis, Tasked Exploration), each of which have unique, sequential actions. This study shows the potential value of unsupervised ML for understanding which types of interactions with simulations could facilitate students’ understanding of complex scientific phenomena.

[1]  R. Millar,et al.  Reasoning from data: How students collect and interpret data in science investigations , 2004 .

[2]  Cristina Conati,et al.  Discovering and Recognizing Student Interaction Patterns in Exploratory Learning Environments , 2010, Intelligent Tutoring Systems.

[3]  P. Deb.  Finite Mixture Models , 2008 .

[4]  P. Shawn Irvin,et al.  Student learning in science simulations: Design features that promote learning gains† , 2011 .

[5]  Cristina Conati,et al.  Combining Unsupervised and Supervised Classification to Build User Models for Exploratory , 2009, EDM 2009.

[6]  Ngss Lead States.  Next generation science standards : for states, by states , 2013 .

[7]  Mohammed J. Zaki,et al.  SPADE: An Efficient Algorithm for Mining Frequent Sequences , 2004, Machine Learning.

[8]  Volodymyr Melnykov,et al.  Finite mixture models and model-based clustering , 2010 .

[9]  Ryan S. Baker,et al.  Educational Data Mining and Learning Analytics , 2014 .

[10]  M. Stieff.  Improving Representational Competence Using Molecular Simulations Embedded in Inquiry Activities , 2011 .

[11]  Sara Khalid,et al.  Machine Learning for Feature Selection and Cluster Analysis in Drug Utilisation Research , 2019, Current Epidemiology Reports.

[12]  Yan Wang,et al.  Research Article Investigating the Effectiveness of Computer Simulations for Chemistry Learning , 2012 .

[13]  M. Linn,et al.  Investigations of a complex, realistic task: Intentional, unsystematic, and exhaustive experimenters , 2011 .

[14]  Janice D. Gobert,et al.  Leveraging Educational Data Mining for Real-time Performance Assessment of Scientific Inquiry Skills within Microworlds , 2012, EDM 2012.

[15]  Hong Jiao,et al.  Data Mining Techniques in Analyzing Process Data: A Didactic , 2018, Front. Psychol..