Spatio-temporal Analysis of Reverted Wikipedia Edits Johannes Kiesel , Martin Potthast, Matthias Hagen, Benno Stein < first name > . < last name > @uni-weimar.de Bauhaus-Universität Weimar www.webis.de ICWSM-17, May 18 th 2017 1 @KieselJohannes
What is Vandalism in Wikipedia? → 2 @KieselJohannes
What is Vandalism in Wikipedia? → 3 @KieselJohannes
Is Vandalism a Problem for Wikipedia? ❑ 470 million article edits to the English Wikipedia in 2003 – 2016 ❑ 40 million (9.5%) are vandalism → a vandalism case every 10s 4 @KieselJohannes
Is Vandalism a Problem for Wikipedia? ❑ 470 million article edits to the English Wikipedia in 2003 – 2016 ❑ 40 million (9.5%) are vandalism → a vandalism case every 10s Countermeasure: Bots that detect and revert vandalism Problem: False positives of the bots discourage editors 5 @KieselJohannes
Towards Understanding Vandalism in Wikipedia How to avoid vandalism in the first place? → Understand why people vandalize Wikipedia. → Analyze when people vandalize. → Analyze where these people are. 6 @KieselJohannes
Towards Understanding Vandalism in Wikipedia How to avoid vandalism in the first place? → Understand why people vandalize Wikipedia. → Analyze when people vandalize. → Analyze where these people are. We analyzed all 1.2 billion edits to the 7 most-edited Wikipedias ❑ Large-scale mining of vandalism using reverted edits ❑ Historical geolocation of anonymous editors by cross-checking several geolocation sources ❑ Spatio-temporal analysis revealing when anonymous editors vandalize from where 7 @KieselJohannes
Mining Vandalism Using Reverted Edits Editor Article revision Edit 8 @KieselJohannes
Mining Vandalism Using Reverted Edits Article over time Editor Article revision Edit 9 @KieselJohannes
Mining Vandalism Using Reverted Edits Article over time Editor ! ! ! Article revision Edit 10 @KieselJohannes
Mining Vandalism Using Reverted Edits Article over time Editor ! ! ! Article revision Edit Reverted edits Revert 11 @KieselJohannes
Mining Vandalism Using Reverted Edits ❑ Not all reverted edits are vandalism ❑ Relying on non-obligatory revert comments underestimates vandalism Identified 7 patterns of non-vandalism or ambiguous reverts Revert to blank page Empty revert Revert correction (enlargement) Revert reverting more than one editor * * * * * * ! ! ! ! ! ! ! ! ! ! Self-revert Reverted revert Interleaved reverts (edit war) + * * * * * ! ! ! ! ! ! ! ! ! ! ! Filter 67% of reverted edits Vandalism detection with precision 82.8%, recall 84.7% 12 @KieselJohannes
Analyzing Vandalism in Wikipedia (by time) 7 edits 6 5 Edits (in millions) 4 3 2 vandalism edits 1 0 6 8 10 12 14 16 18 20 22 0 2 4 Hour of day 13 @KieselJohannes
Analyzing Vandalism in Wikipedia (by time) 0.5 0.4 vandalism edits vandalism ratio = edits Vandalism ratio 0.3 0.2 0.1 0.0 6 8 10 12 14 16 18 20 22 0 2 4 Hour of day 14 @KieselJohannes
Analyzing Vandalism in Wikipedia (by time) 0.5 English Wikipedia from United States 0.4 Monday - Thursday Friday Vandalism ratio 0.3 Saturday Sunday 0.2 Estimates from less than 1000 vandalism 0.1 edits are shown as dotted lines 0.0 6 8 10 12 14 16 18 20 22 0 2 4 Hour of day 15 @KieselJohannes
Analyzing Vandalism in Wikipedia (by time) 0.5 0.5 English Wikipedia English Wikipedia from United States from Canada 0.4 0.4 Vandalism ratio Vandalism ratio 0.3 0.3 Monday - Thursday 0.2 0.2 Friday 0.1 0.1 Saturday 0.0 0.0 Sunday 6 8 10 12 14 16 18 20 22 0 2 4 6 8 10 12 14 16 18 20 22 0 2 4 Hour of day Hour of day Estimates from less 0.5 0.5 English Wikipedia English Wikipedia from Australia from United Kingdom than 1000 vandalism 0.4 0.4 edits are shown as Vandalism ratio Vandalism ratio 0.3 0.3 dotted lines 0.2 0.2 0.1 0.1 0.0 0.0 6 8 10 12 14 16 18 20 22 0 2 4 6 8 10 12 14 16 18 20 22 0 2 4 Hour of day Hour of day 16 @KieselJohannes
Analyzing Vandalism in Wikipedia (by time) 0.5 Japanese Wikipedia from Japan 0.4 Monday - Thursday Friday Vandalism ratio 0.3 Saturday Sunday 0.2 Estimates from less than 1000 vandalism 0.1 edits are shown as dotted lines 0.0 6 8 10 12 14 16 18 20 22 0 2 4 Hour of day 17 @KieselJohannes
Analyzing Vandalism in Wikipedia (by time) 0.5 French Wikipedia from France 0.4 Monday - Thursday Friday Vandalism ratio 0.3 Saturday Sunday 0.2 Wednesday Estimates from less than 1000 vandalism 0.1 edits are shown as dotted lines 0.0 6 8 10 12 14 16 18 20 22 0 2 4 Hour of day 18 @KieselJohannes
Analyzing Vandalism in Wikipedia (by country) Vandalism ratio 0.12 0.16 0.2 0.24 0.28 Country estimates from less than 1000 vandalism edits are not colored 19 @KieselJohannes
Analyzing Vandalism in Wikipedia (by country with English as an official language) Vandalism ratio 0.12 0.16 0.2 0.24 0.28 Country estimates from less than 1000 vandalism edits are not colored 20 @KieselJohannes
Analyzing Vandalism in Wikipedia (by time) 0.5 0.5 English Wikipedia German Wikipedia from Germany from Germany 0.4 0.4 Vandalism ratio Vandalism ratio 0.3 0.3 0.2 0.2 0.1 0.1 0.0 0.0 6 8 10 12 14 16 18 20 22 0 2 4 6 8 10 12 14 16 18 20 22 0 2 4 Hour of day Hour of day Monday - Thursday Friday Saturday Sunday Estimates from less than 1000 vandalism edits are shown as dotted lines 21 @KieselJohannes
Analyzing Vandalism in Wikipedia (by time) 0.5 English Wikipedia German Wikipedia from Germany from Germany 0.4 Vandalism ratio Vandalism ratio 0.3 0.1 0.2 0.1 0.0 0.0 6 8 10 12 14 16 18 20 22 0 2 4 6 8 10 12 14 16 18 20 22 0 2 4 Hour of day Hour of day Monday - Thursday Friday Saturday Sunday Estimates from less than 1000 vandalism edits are shown as dotted lines 22 @KieselJohannes
Spatio-temporal Analysis of Reverted Wikipedia Edits Future Work ❑ Identify different types of vandalism ❑ Identify changes in vandalism behavior over the years Resources ❑ Interactive tool for exploring the vandalism ratio graphs webis16.medien.uni-weimar.de/wikipedia-vandalism ❑ Supplementary material ( ∼ 50 pages of tables and graphs) github.com/webis-de/ICWSM-17/raw/master/supplementary-material.pdf ❑ Code for historical geolocation github.com/webis-de/aitools4-aq-geolocation ❑ Code for reproducing experiments github.com/webis-de/ICWSM-17 23 @KieselJohannes
Recommend
More recommend
