A new exploratory data mining technique for identifying important student learning behaviors and strategies is grabbing entrepreneurial interest and kudos from the international community.

John Kinnebrew, a research associate at Vanderbilt’s Institute for Software Integrated Systems, and Professor Gautam Biswas in the Electrical Engineering and Computer Science Department, received the “Best Paper Award” at the recent International Conference on Educational Data Mining (EDM 2012) in Chania, Greece.

Their paper, “Identifying Learning Behaviors by Contextualizing Differential Sequence Mining with Action Features and Performance Evolution,” presents how the new technique achieves results through the analysis of the complex data generated by computer-based learning environments.

“We are pleased that the paper won the EDM award,” Biswas said, “and we are also excited to be discussing commercial applications of the technique with two private companies.”

Data mining uses artificial intelligence techniques, neural networks, and advanced statistical tools to reveal trends, patterns, and relationships, which might otherwise have remained undetected.

To assess and compare students’ learning behaviors, the new method used by Biswas and Kinnebrew extends sequence mining algorithms with a variety of context information, such as changes in task performance during learning. “Specifically, a piecewise linear segmentation algorithm is first applied to student performance data from the learning environment to identify segments of students’ productive and unproductive learning activity,” Biswas explained.

These segments are then compared using a combination of action abstraction and sequence mining algorithms to identify important learning behaviors that differ between productive and unproductive activity segments or between student groups.

“The goal of this research is to leverage the wealth of data tracked by computer-based learning environments in order to assess, model, and understand student learning behaviors and strategies more accurately,” said Kinnebrew. “Ultimately, this adaptive strategy support allows computer-based learning environments to enhance student learning, not only for a particular curriculum topic, but also in preparation for future learning, by promoting more effective learning strategies and self-regulation.”

The innovative methods recognized by the EDM award build on successful approaches applied by Kinnebrew and Biswas in several different projects at Vanderbilt, including Betty’s Brain, an animated computer learning program developed for middle school science students, SimSelf, and other game environments for teaching students.

Biswas earned his B.Tech. degree in electrical engineering from the Indian Institute of Technology in Bombay, and M.S. and Ph.D. degrees in computer science from Michigan State University. Kinnebrew has a B.S. degree in computer science from Harvard University and his M.S. and Ph.D. degrees from Vanderbilt University.

Posted on Thursday, September 6, 2012 in Alumni, Electrical Engineering and Computer Science, News, Research.