Порядковая оценка популярности публикаций для задач мониторинга онлайновых социальных медиа

Abstract

The paper examines the possibility of using various methods for generating ranking estimates of the popularity of messages in online social media. Using experimental data collected from the VKontakte network as an example, it is shown that using quantitative indicators of likes, reposts and comments as the basis for an ranking assessment of popularity gives an equivalent result. The need to use popularity metrics, normalized by the number of views or the size of the community audience, is noted when analyzing several online communities at the same time.