This is a pre-production deployment of Warehouse, however changes made here WILL affect the production instance of PyPI.
Latest Version Dependencies status unknown Test status unknown Test coverage unknown
Project Description

The edxclassify package contains a classification suite built to detect learner affect and behavior in the discussion forums of Massive Open Online Courses (MOOCs). It is a result of a year’s worth of research at Stanford and was designed to enable the discovery of insights into forums attached to Stanford’s online course offerings.

Research and Motivation

Our motivation to discover insights into the dynamics of these courses is three-fold. In particular, we wish to:

  1. better understand the educational and learning process,
  2. improve the educational environment for MOOC learners, and
  3. empower instructors by equipping them with a suite of tools designed to ease the burden of teaching large-scale classes.

For example, we could use a classifier that detects confusion in forum posts to help us target automated learning interventions in courses. We have built such a prototype; to learn more, please refer to our paper, published in the eighth conference on Educational Data Mining.

Included Classifiers

The abstractions in edxclassify are general enough to be applicable to most classification tasks. The repository does come packaged with a set of classifiers that were trained to detect affect in Stanford’s MOOC discussion forums. Since Stanford’s courses are powered by edX, these classifiers should be compatible with any edX MOOC; they should also be compatible with other flavors of MOOCs, as their feature space is not tightly coupled with the particulars of edX.

These classifiers were trained on subsets of the Stanford MOOC-Posts Dataset, a collection of 30,000 human-tagged forum posts, originating from a variety of courses. Classifiers to detect all six of the core variables in the MOOC-Posts Dataset – confusion, urgency, sentiment, question, answer, and opinion – are included in this repository. The class edxclassify.live_clf.LiveCLF provides an interface to them; see the module edxclassify.live_clf for further documentation.

Running Experiments

edxclassify.harness is a driver that facilitates the training and testing of and experimentation with classifiers. After installation, it can be invoked with the command clfharness.

Data

The MOOC-Posts Dataset is available to researchers, upon request.

Much of the included code in edxclassify was designed for data formatted as per edxclassify.feature_spec; in particular, the harness takes data files each containing a pickled list of examples, each example a list with features in the positions specified in edxclassify.feature_spec. If you would like access to these data files, first request access to the MOOC-Posts Dataset. When your request is approved, send an email to akshayka ~at~ cs.stanford.edu with subject line “edxclassify: request for data files”.

Installation

Installation can proceed in two ways: from source or from pip. Note that when installing from pip, only a subset of the pre-trained classifiers found in this repository will be included, due to size constraints imposed by pypi. In particular, the pypi version only includes classifiers for confusion, whereas the source version includes classifiers for all six MOOC-Posts variables.

Regardless of whether you install from source or from pip, begin by installing scikit-learn and its dependencies; make sure to install version 0.15.2 to ensure compatibility with skll.

If installing from source, clone this repository and simply run python setup.py install. Otherwise, run pip install edxclassify.

Release History

Release History

0.10.a1

This version

History Node

TODO: Figure out how to actually get changelog content.

Changelog content for this version goes here.

Donec et mollis dolor. Praesent et diam eget libero egestas mattis sit amet vitae augue. Nam tincidunt congue enim, ut porta lorem lacinia consectetur. Donec ut libero sed arcu vehicula ultricies a non tortor. Lorem ipsum dolor sit amet, consectetur adipiscing elit.

Show More

Download Files

Download Files

TODO: Brief introduction on what you do with files - including link to relevant help section.

File Name & Checksum SHA256 Checksum Help Version File Type Upload Date
edxclassify-0.10.a1.tar.gz (10.7 MB) Copy SHA256 Checksum SHA256 Source Jun 17, 2015

Supported By

WebFaction WebFaction Technical Writing Elastic Elastic Search Pingdom Pingdom Monitoring Dyn Dyn DNS HPE HPE Development Sentry Sentry Error Logging CloudAMQP CloudAMQP RabbitMQ Heroku Heroku PaaS Kabu Creative Kabu Creative UX & Design Fastly Fastly CDN DigiCert DigiCert EV Certificate Rackspace Rackspace Cloud Servers DreamHost DreamHost Log Hosting