Forecasting the presence and intensity of hostility on Instagram using linguistic and social features

In this paper, we propose a method to forecast the arrival of hostile comments on Instagram posts. In order to support different intervention strategies, as well as to assess the difficulty of variants of this problem, we consider two forecasting tasks: (1) hostility presence forecasting: given the initial sequence of non-hostile comments in a post, predict whether some future comment will be hostile; (2) hostility intensity forecasting: given the first hostile comment in a post, predict whether the post will receive more than N hostile comments in the future. Solutions to the first task would support more aggressive interventions that attempt to eliminate all hostile comments from the system, while solutions to the second task would focus interventions on the most extreme cases.

We proposed methods to forecast both the presence and intensity of hostility in Instagram comments. Using a combination of linguistic and social features, the best model produces an AUC of 0.82 for forecasting the presence of hostility ten or more hours in the future, and an AUC of 0.91 for forecasting whether a post will receive more than 10 hostile comments or only one hostile comment. We find several predictors of future hostility, including (1) the post’s author has received hostile comments in the past; (2) the use of user directed profanity; (3) the number of distinct users participating in a conversation; and (4) trends in hostility thus far in the conversation.