Submissions/WikiQuiz - automatic generation of multiple choice question sets using the Wikimedia projects, raw text processing and metadata tags

From Wikimania 2010 • Gdańsk, Poland • July 9-11, 2010
Jump to: navigation, search


Information

This is an open submission for Wikimania 2010.


Title of the submission
WikiQuiz - automatic generation of multiple choice question sets using the Wikimedia projects, raw text processing and metadata tags
Type of submission (workshop, tutorial, panel, presentation)
Presentation
Author of the submission
Kristoffer Mellberg
E-mail address or username (if username, please confirm email address in Special:Preferences)
kristoffer dot mellberg at gmail dot com
Country of origin
Sweden
Affiliation, if any (organization, company etc.)
Wikimedia Sverige
Personal homepage or blog
Abstract (please use no less than 300 words to describe your proposal)

WikiQuiz is an idea for a system consisting of three parts - a question/answers set generation subsystem, a quality improvement subsystem for the existing quiz sets and a quiz game using the former and giving feedback to the latter.

The subsystem for creating sets of one question a few answer alternatives would have several mechanisms for building these sets. One would probably be a REGEXP-style search for sentences in Wikipedia articles that appear to state that some entity has a characteristic which only one entity can have, for instance being the highest mountain in a certain region. The mechanism would then use the name of the article (or the DEFAULTSORT value of the article) as the right answer and find alternative answers by looking in the categories the artices belongs to. Other mechanism would use more extensive use of category tags and similar machine-readable information to create questions along the lines of "which of the following is _not_ a X", "which X died in the year Y". Images from Wikimedia Commons could be used to illustrate some questions, perhaps allowing several difficulty levels where the image is first shown through a blur or pixelation filter of varying intensity.

The quality subsystem would keep track on how many times a question/answer set had been used, what percentage of these times it had been answered correctly and perhaps some sort of user rating, either implemented as a simple "this question does not make sense at all" feedback mechanism or as a more complex rating.

Lastly, the quiz game could be implemented in many ways, and perhaps there could be many different games (alone, for several different players at the same site, for several players in different locations...) for different platforms (standalone applications for ordinary computers, web versions, embedded applications for Facebook and similar sites, smartphone apps, video game console games). These games could either use a downloaded set of questions and answers or stream the questions, but on-line versions would have the advantage of being able to offer a link to the Wikipedia articles used when generating the question and answers, making it possible for the user not only to find out what the right answer was but also to learn more.

Track (People and Community/Knowledge and Collaboration/Infrastructure)

Knowledge and collaboration, perhaps?

Will you attend Wikimania if your submission is not accepted?

Yes.

Slides or further information (optional)


Interested attendees

If you are interested in attending this session, please sign with your username below. This will help reviewers to decide which sessions are of high interest. Sign with four tildes. (~~~~).

  1. Add your username here.