Birkbeck University of London Knowledge Lab | Events | Exploiting temporal features for efficient classification of educational data

Document Actions

Exploiting temporal features for efficient classification of educational data

Nov 19, 2015
When Nov 19, 2015
from 01:00 PM to 02:00 PM
Where Birkbeck Main Building, Room 151
Attendees Professor Panos Papapetrou, Stockholm University
Add event to calendar vCal

In many application domains, including educational data mining, users are characterised by time evolving features.  For example, students participating in a course project perform a sequence of communication activities with other group members.  Such communication events and their temporal order may reveal the role of each student within the project. In this talk, I will describe a novel sequence classification framework that generates features based on frequent patterns at multiple levels of time granularity. Feature selection techniques are applied to identify the most informative features that are then used to construct the classification model.  Experiments have been performed on data collected from an asynchronous communication tool in which students interact to accomplish a group project. The experimental results showed that the model can achieve competitive performance in detecting the students' roles, compared to a baseline similarity-based approach.  Finally, the second part of my talk will briefly focus on Goldeneye, a recent method for evaluating the performance of different classifiers and learning groupings of features based on some predictive performance quality measure. Implications for potential application to educational data will be discussed.


Panagiotis Papapetrou is associate professor at the Department of Computer and Systems Sciences at Stockholm University. He is also adjunct professor at Aalto University, Finland. His area of expertise is algorithmic data mining with particular focus on mining and indexing sequential data, mining  complex metric and nonmetric spaces, biological sequences, time series, and sequences of interval-based events. Panagiotis earned his PhD in computer science at Boston University in 2009, was a post-doctoral researcher at Aalto University during 2009-2013 and lecturer at Birkbeck University of London during 2012-2013. He has participated in 4 EU Projects, 2 NSF projects, and 2 Academy of Finland.

Filed under: