Selecting the Most Probable Author of Asocial Posting in Online Media
Kristína Machová, P. Staronova · 2019
The paper is focused on the field of the detection of antisocial behavior in online communities. Within this field, the problem of an assigning of an author to an unknown suspicious document was solved. This problem solving supposes that a set of short informal documents posts of a set of known authors of the troll type are at disposal. Presented method of authorship attribution is based on the recognition of special dictionary of suspicious author. The designed method was implemented. It can be used to select the most probable author of toxic posts. The paper describes also tests of the designed method effectivity on short documents and a comparison with related works.