Beyond Tutor Logs: Using Sensors to Augment Intelligent Tutoring System Logs 
David Cooper 

The session will start with an overview of sensors (cameras, EEG, eye trackers, pressure sensors, skin conductivity, Heart Rate, etc.) that have been used in intelligent tutoring systems and their success. Then, we will move to discuss the challenges of using sensors, including the need for synchronization, calibration, and installation. Next we will discuss some intermediate tools that can be used for handling sensor data such as facial expression recognition (The Computer Emotion Recognition Toolbox (CERT), Constrained Local Models), voice signal processing (PRAAT, MARSYAS, OpenSMILE), and other signal processing (MATLAB, Aquila, BioSig). Then we will discuss summarization of the sensor data in order to add them as additional features of the tutor logs. We will conclude with an example of sensors being used in a math tutoring system called Wayang Outpost. 

Link to tutorial slides

Sequence-based Motif Discovery: automatically finding student patterns of behavior over time 
David Cooper 

The goal of this mini-tutorial is to present the basic steps of sequence-based motif discovery, the reason for using it, and the tools available to implement your own version. We will begin by defining a motif and then discuss potential reasons for searching for patterns over time. This mini-tutorial will describe the basic algorithm and will discuss steps of preprocessing log data. It will include a case study on student patterns using the Wayang Outpost intelligent tutoring system. The motif discovery software is available to the community

Getting Started with RapidMiner 
Agathe Merceron

RapidMiner is a general data mining tool. This tutorial is intended for beginners with RapidMiner. The tutorial will show how to build a data mining process with RapidMiner. In particular the data exploration facilities of RapidMiner will be shown. They can be very helpful to guide the modeling phase or to interpret results given by some operators like k-means.

An open source R toolkit for EDM 
Tristan Nixon

Over the last several years, Carnegie Learning Inc., in conjunction with others, has developed a suite of tools to do various common EDM tasks in the R language. These include fetching data from a Datashop formatted database, plotting learning curves, and fitting and evaluating Bayesian Knowledge Tracing (BKT) models. We are making these tools available to the EDM community, including documentation and source code. This tutorial will walk through the packages, and the functions they provide, and demo how to use these on some example data.