Wikidata:ORES/Vandalism patterns

From Wikidata
Jump to navigation Jump to search

Patterns[edit]

Please list patterns in vandalisms you see. Things like "If it's made by a unregistered user, it's very likely", "Edits on soccer players are more likely to be vandalism", "People vandalise descriptions more often", etc. Thank you! Amir Sarabadani (WMDE) (talk) 15:37, 27 June 2018 (UTC)

Vandalism[edit]

  • A "meta" pattern: edit that introduce constraint violations are more likely to be vandalism. But it is costly to compute all violations so starting for example by "format" and "one of" constraints is probably a good idea (they do not depend on global state). Tpt (talk) 10:10, 2 July 2018 (UTC)
  • Changing only a digit of a number without being the least significant digit. The more significant that digit is, the more likely that edit is to be vandalism: "1784" → "1884". --abián 11:11, 2 July 2018 (UTC)
  • Adding several digits to a number: "520" → "5205462", "520" → "1111520", etc. --abián 11:11, 2 July 2018 (UTC)
  • "Political issues" [1] Matěj Suchánek (talk) 07:30, 3 July 2018 (UTC)
  • Changes to (properly) referenced statements. Matěj Suchánek (talk) 10:37, 19 July 2018 (UTC)

Strings[edit]

  • Insults in the description: "stupid", "fool", etc. However, these names can be valid for labels and aliases. --abián 10:21, 2 July 2018 (UTC)
    ✓ Done This already has been implemented Amir Sarabadani (WMDE) (talk) 14:00, 6 August 2018 (UTC)
  • Strings of repeated characters: "aaaa". --abián 10:27, 2 July 2018 (UTC)
    ✓ Done This already has been implemented Amir Sarabadani (WMDE) (talk) 14:00, 6 August 2018 (UTC)
  • Strings of repeated pairs of characters: "ioioioio". --abián 10:43, 2 July 2018 (UTC)
  • Many consecutive characters in capital letters. --abián 10:33, 2 July 2018 (UTC)
    ✓ Done Added Amir Sarabadani (WMDE) (talk) 14:00, 6 August 2018 (UTC)
  • Too many consecutive vowels or consonants. --abián 10:33, 2 July 2018 (UTC)
  • Adding letters to an already existing string without introducing a blank space: "Douglas Adams" → "Douglas Adamstyui". --abián 10:33, 2 July 2018 (UTC)
  • Replacing a label that is repeated in more languages with a very different one that isn't present in any other language: {"Douglas Adams"@en, "Douglas Adams"@es, "Douglas Adams"@de, "Douglas Adams"@fr} → {"Jimbo Wales"@en, "Douglas Adams"@es, "Douglas Adams"@de, "Douglas Adams"@fr}. --abián 10:38, 2 July 2018 (UTC)
  • Strings made up by pressing adjacent keys: "asdf", "sdfg", "dfgh", etc. --abián 10:42, 2 July 2018 (UTC)
  • Descriptions containing "wikid" or "wikip" for items that aren't related to the Wikimedia movement. --abián 10:59, 2 July 2018 (UTC)
  • Complete URLs (including "http://" or "https://"). --abián 10:59, 2 July 2018 (UTC)
    ✓ Done Added Amir Sarabadani (WMDE) (talk) 14:00, 6 August 2018 (UTC)
  • Descriptions containing imperative forms or orders: "suck it", "make it", "f*ck", "don't be", "don't piss", etc. --abián 11:21, 2 July 2018 (UTC)
    ✓ Done Added Amir Sarabadani (WMDE) (talk) 14:00, 6 August 2018 (UTC)
  • Descriptions in the first or second person: "my", "me", "I", "mine" "your", "you", "yours", etc. --abián 11:21, 2 July 2018 (UTC)
    ✓ Done Added Amir Sarabadani (WMDE) (talk) 14:00, 6 August 2018 (UTC)
  • User:Pasleim/Vandalism. Matěj Suchánek (talk) 10:33, 19 July 2018 (UTC)

Edits in good faith[edit]

  • My edits. :P --abián 10:21, 2 July 2018 (UTC)
  • Undoing edits that meet vandalism patterns. --abián 10:27, 2 July 2018 (UTC)