Emotional Analysis of posts in VKontakte: classification or regression?
А. В. Колмогорова, A. A. KALININ · Computational Linguistics and Intellectual Technologies · 2022
The article summarizes the results of two tasks in machine learning paradigm: the task of classification according to the criterion of dominating emotion on the data of social networks posts in Russian and the regression task using the same data. The experiments are conducted on the data set collected from VKontakte social network and consisted of 3879 posts assessed by 2000 informants on Toloka crowd sourcing platform. The annotation procedure was based on the original interface for non-discrete emotion assessment elaborated by researchers.