Дослідження методів семантичної кластеризації для аналізу новин

Abstract

This work is dedicated to clustering news based on their topics and exploring methods of semantic clustering. Semantic clustering can be used to increase the number of matches and differences between text elements. For advanced semantic clustering, machine learning methods, such as cluster analysis, clustering algorithms, or neural measures that capture semantic interactions between objects, can be used. An analysis of the most popular methods of semantic clustering is conducted, along with testing their effectiveness and speed of operation. Nevertheless, remaining trends in advanced semantic clustering of text documents include the emergence of current models of deep innovation and hybrid approaches