Deep Learning Models That Predict Conflicts In Online Communities - Abhijeet Katte, Analytics India Magazine

There is not enough research and knowledge around how interactions happen online between communities and users, especially in the space of conflicts. A team from the Computer Science and Linguistics department of Stanford University wanted to change just that. What they did: research conflict events by searching for cases where one community posted a hyperlink to another community. What they found: conflicts tend to be initiated by a handful of communities—less than 1% of communities start 74% of conflicts. In the long term, conflicts have adverse effects and reduce the overall activity of users in targeted communities. The researchers also came up with a way to predict conflicts on the web communities. https://analyticsindiamag.com/deep-learning-models-that-predicts-conflicts-in-online-communities/