Jump to content

Proposal:Algorithm to assign an estimated credibility to articles

From Strategic Planning
Status (see valid statuses)

The status of this proposal is:
Request for Discussion / Sign-Ups

Every proposal should be tied to one of the strategic priorities below.

Edit this page to help identify the priorities related to this proposal!


  1. Achieve continued growth in readership
  2. Focus on quality content
  3. Increase Participation
  4. Stabilize and improve the infrastructure
  5. Encourage Innovation


Summary

Many times when you look at a wikipedia article you can't tell whether is credible or not. There should be an algorithm to give it an "estimated" credibility.

Proposal

Let's take an article about some disease. It's VERY important there is credible info as a person's life *potentially* could be on the line.

Of course you can't compute credibility perfectly but there are generally some good hints

  • Verified Qualified Contributors: for example, if a verified MD contributes to an article about a certain disease it would have an impact on the credibility
  • Unverified Qualified Contributors: say somebody writes articles about all kinds of diseases and so far has not gotten any rollbacks, etc. Chances are his articles are kinda good
  • References: this gets a bit tricky. How can you verify a reference. Well there are few ways. First of all, just because an article quotes a good article (ex: from a respected Medical journal) doesn't mean it was correctly quoted. That's why there should be a regex engine analyzing quoted sources for validity of information (ex: if nausea is listed under symptoms, regex engine checks for that word and checks around for a heading named symptom. If multiple sources are quoted, even if not for specific symptom, regex engine will check words "around" that area against all sources
  • "Valid" Stamp: if a person of credibility on the subject reviews an article, he can give a "stamp" on a version. Users may opt to see the latest "stamped" version of an article
  • Regex related articles (ex: regex article on nausea for that disease)
  • Regex Google results
  • Disclaimers: if the estimated credibility is less than a certain standard or if the potential severity of said disease is large, there should be a disclaimer saying the information may not be accurate and to visit a doctor if in doubt.

NOTE: The same process can be applied for just about any time of article. I just used diseases as an example because I just got through 10 episodes of House in a row XD Adz 23:46, 16 August 2009 (UTC)

Motivation

  • If done correctly, it can yield an amazing credibility checking system that can be applied virtually anywhere.
  • Can yield credibility that can exceed that of encyclopedias
  • Credibility engine may be used for research

Key Questions

  • What *really* matters when it comes to credibility

Potential Costs

  • Obviously the processing power. Regex is VERY heavy on the processor, especially on such a detailed scale
  • Mathematicians, statisticians, programmers
  • Way to verify a person's identity without encroaching on his/her privacy (unfortunately trust and credibility don't go well together :()
  • Massive amounts of research
  • Experts in certain fields (ex: doctors, physicists, etc) for tests of the system.

References

Community Discussion

Do you have a thought about this proposal? A suggestion? Discuss this proposal by going to Proposal talk:Algorithm to assign an estimated credibility to articles.

Want to work on this proposal?

  1. Vibhijain 11:02, 6 May 2011 (UTC)
  2. .. Sign your name here!