Day: June 1, 2015

Organized by e-learning Pedagogical Support Unit, CETL

Abstract:

Understanding why students stopout will help in understanding how students learn in Massive Open Online Courses (MOOCs). In this seminar, Dr. Una-May O’Reilly will describe how she and her research group build accurate predictive models of MOOC student stopout via a scalable, prediction methodology, end to end, from raw source data to model analysis. They attempted to predict stopout for the Fall 2012 offering of MIT’s 6.002x.

This involved the meticulous and crowd-sourced engineering of over 25 predictive features extracted for thousands of students, the creation of temporal and non-temporal data representations for use in predictive modeling, the derivation of over 10 thousand models with a variety of state-of-the-art machine learning techniques and the analysis of feature importance by examining over 70,000 models. They found that stopout prediction is a tractable problem. Their models achieved an AUC (receiver operating characteristic area-under-the-curve) as high as 0.95 (and generally 0.88) when predicting one week in advance. Even with more difficult prediction problems, such as predicting stop out at the end of the course with only one weeks’ data, the models attained AUCs of ~0.7.

About the Speakers:

Dr. Una-May O’Reilly (http://people.csail.mit.edu/unamay/) leads the AnyScale Learning For All (ALFA) group (http://groups.csail.mit.edu/ALFA) at Massachusetts Institute of Technology Computer Science and Artificial Intelligence Laboratory. ALFA focuses on scalable machine learning, evolutionary algorithms, and frameworks for knowledge mining, prediction and analytics. She received the EvoStar Award for Outstanding Achievements in Evolutionary Computation in Europe in 2013 and serves as Vice-Chair of ACM Special Interest Group for Genetic and Evolutionary Computation (SIGEVO).

Miss Carmen Cheung
Email: carmen.cheung@hku.hk.