Розв’язання задачі кластеризації наукових текстів

Abstract

This study delves into the formal and substantive aspects of the problem of thematic modeling of scientific texts through the division of probability distribution mixtures. The key objective is to identify the subject of each document from a collection of scientific documents. The collection is comprised of textual scientific documents sourced from a variety of places. The simulation will yield a set of documents, with each marked by multiple topics from a general list. These topics will each be associated with a set of keywords that best represent them.