Clustering of social media content with the use of BigData technology

Abstract

This work is devoted to one of the key problems arising in the analysis of social media - the problem of account classification on the basis of media uploaded by users. The main difficulties in solving the problem are the heterogeneous nature of the content (photos, artworks, greeting cards, etc.) and colossal volumes of analyzed information, which leads to excessive computational complexity of its processing. In the paper, we discuss an approach to social media clustering based on class annotation, using BigData technology - a modern and effective tool to handle the described difficulties. To carry out computational experiments, we collected a large sample of images from real profiles of Twitter users.