AntiVandalismFilter

Last modified by Sergiu Dumitriu on 2008/03/28 00:16

Project name
Anti Vandalism Filters
Mentor(s)
Estimated workload
2 man month
Details

While access rights are a global way of ensuring that only trusted people can edit, and CAPTCHA gives moderate results for preventing automatic spamming, an automatic vandalism detection mechanism is better as it allows anybody to edit, and does not annoy people with deciphering terribly distorted letters.

Wikipedia does that with the help of bots that revert suspect edits. XWiki can block suspect edits before they even happen.

Some filters/features:

  • Autodetect clients making mass requests and block their IP for a certain amount of time
  • Autodetect bad edits based on the entered text (machine learning algorithms)
  • Work with an anti-spam service that lists "bad" IPs
  • Users can indicate bad edits so that the filter can learn from new instances
Importance
3
Year

2008

Developer profile

Required:

  • Good Java programming
  • Machine Learning
  • Algorithms and Complexity

Helps:

  • Familiar with the XWiki code
  • Design Patterns
Tags:
Created by Sergiu Dumitriu on 2008/03/21 16:33
   

Get Connected