Conversational Networks for Automatic Online Moderation
Abstract
Moderation of user-generated content in an online community is a challenge that has great socio-economical ramifications. However, the costs incurred by delegating this work to human agents are high. For this reason, an automatic system able to detect abuse in user-generated content is of great interest. There are a number of ways to tackle this problem, but the most commonly seen in practice are word filtering or regular expression matching. The main limitations are their vulnerability to intentional obfuscation on the part of the users, and their context-insensitive nature. Moreover, they are language-dependent and may require appropriate corpora for training. In this paper, we propose a system for automatic abuse detection that completely disregards message content. We first extract a conversational network from raw chat logs and characterize it through topological measures. We then use these as features to train a classifier on our abuse detection task. We thoroughly assess our system on a dataset of user comments originating from a French Massively Multiplayer Online Game. We identify the most appropriate network extraction parameters and discuss the discriminative power of our features, relatively to their topological and temporal nature. Our method reaches an F-measure of 83.89 when using the full feature set, improving on existing approaches. With a selection of the most discriminative features, we dramatically cut computing time while retaining most of the performance (82.65).
Fichier principal
main.pdf (1.53 Mo)
Télécharger le fichier
abuse_distro.pdf (78.89 Ko)
Télécharger le fichier
after_1.pdf (4.69 Ko)
Télécharger le fichier
before_1.pdf (4.85 Ko)
Télécharger le fichier
collisions.pdf (119.9 Ko)
Télécharger le fichier
comp.bib (56.84 Ko)
Télécharger le fichier
distributions.pdf (84.54 Ko)
Télécharger le fichier
fig_concepts.tex (1.77 Ko)
Télécharger le fichier
fig_graph_build.tex (2.68 Ko)
Télécharger le fichier
full_1.pdf (5.07 Ko)
Télécharger le fichier
photo_dufour.pdf (327.66 Ko)
Télécharger le fichier
photo_labatut.png (321.96 Ko)
Télécharger le fichier
photo_linares.png (36.16 Ko)
Télécharger le fichier
photo_papegnies.png (223.28 Ko)
Télécharger le fichier
res_context_size.pdf (13.21 Ko)
Télécharger le fichier
res_full_vs_before_vs_after_vs_all.pdf (15.61 Ko)
Télécharger le fichier
res_window_size.pdf (16.04 Ko)
Télécharger le fichier
uu_ud_wu_wd.pdf (15.52 Ko)
Télécharger le fichier
Origin | Files produced by the author(s) |
---|
Origin | Files produced by the author(s) |
---|
Origin | Files produced by the author(s) |
---|
Origin | Files produced by the author(s) |
---|