TY - JOUR

T1 - Hierarchical Classification of Scientific Articles Using Deep Learning (Using the UDC Hierarchy As an Example)

AU - Мамедов, Валентин Юрьевич

AU - Ковалевский, Данил Анатольевич

AU - Морозов, Дмитрий Алексеевич

AU - Столяров, Степан Сергеевич

AU - Оспичев, Сергей Сергеевич

N1 - Mamedov, V.Y., Kovalevsky, D.A., Morozov, D.A. et al. Hierarchical Classification of Scientific Articles Using Deep Learning (Using the UDC Hierarchy As an Example). Aut. Control Comp. Sci. 59, 1181–1192 (2025). https://doi.org/10.3103/S0146411625700440

PY - 2025/12

Y1 - 2025/12

N2 - The exponential growth of scientific publications has heightened the need for robust tools to organize and retrieve research effectively. The Universal Decimal Classification (UDC) serves as a valuable framework for categorizing articles by subject area. However, manual assignment of UDC codes is often prone to inaccuracies or oversimplification, limiting its utility. In this study, we present a novel approach for the automated assignment of UDC codes to scientific articles using BERT-based models. Our methodology is trained and evaluated on a dataset comprising over 19 000 articles in mathematics and related disciplines. To address the hierarchical structure of the UDC, we develop two specialized evaluation metrics: hierarchical classification accuracy and hierarchical recommendation accuracy. We also explore multiple strategies for flattening hierarchical labels. Our results demonstrate a hierarchical recommendation accuracy of 0.8220. Furthermore, blind expert evaluation reveals that discrepancies between the reference and predicted labels often stem from errors in the original UDC code assignments by the authors of articles. Our approach demonstrates strong potential for automating the classification of scientific articles and can be extended to other hierarchical classification systems.

AB - The exponential growth of scientific publications has heightened the need for robust tools to organize and retrieve research effectively. The Universal Decimal Classification (UDC) serves as a valuable framework for categorizing articles by subject area. However, manual assignment of UDC codes is often prone to inaccuracies or oversimplification, limiting its utility. In this study, we present a novel approach for the automated assignment of UDC codes to scientific articles using BERT-based models. Our methodology is trained and evaluated on a dataset comprising over 19 000 articles in mathematics and related disciplines. To address the hierarchical structure of the UDC, we develop two specialized evaluation metrics: hierarchical classification accuracy and hierarchical recommendation accuracy. We also explore multiple strategies for flattening hierarchical labels. Our results demonstrate a hierarchical recommendation accuracy of 0.8220. Furthermore, blind expert evaluation reveals that discrepancies between the reference and predicted labels often stem from errors in the original UDC code assignments by the authors of articles. Our approach demonstrates strong potential for automating the classification of scientific articles and can be extended to other hierarchical classification systems.

KW - КЛАССИФИКАЦИЯ ТЕКСТОВ

KW - ИЕРАРХИЧЕСКАЯ КЛАССИФИКАЦИЯ ТЕКСТОВ

KW - УНИВЕРСАЛЬНЫЙ ДЕСЯТИЧНЫЙ КЛАССИФИКАТОР

KW - ГЛУБОКОЕ ОБУЧЕНИЕ

KW - TEXT CLASSIFICATION

KW - HIERARCHICAL TEXT CLASSIFICATION

KW - UNIVERSAL DECIMAL CLASSIFIER

KW - DEEP LEARNING

UR - https://www.scopus.com/pages/publications/105030608969

UR - https://elibrary.ru/item.asp?id=80479012

U2 - 10.3103/S0146411625700440

DO - 10.3103/S0146411625700440

M3 - Article

VL - 59

SP - 1181

EP - 1192

JO - Automatic Control and Computer Sciences

JF - Automatic Control and Computer Sciences

SN - 1558-108X

IS - 7

ER -