Ensemble Learning in the Recognition of Various Types of the Online Toxicity

This article offers approaches to classifying online toxicity using a set of methods learning. It focuses on detecting toxic posts mainly on fake news and offensive posts. The article also provides definitions of basic terms, problems associated with recognizing toxic posts, places on the web where they occur, as well as specific steps that can be taken to protect users from them. Based on the analysis of the current state and the latest knowledge, various methods of ensemble machine learning are proposed, which have been used for training, testing and evaluation on different groups of toxic posts. The article also contains detailed test results.