Алгоритм именования кластеров при автоматизированном формировании критериев оценивания программного продукта

Никита Андреевич Экерт; Ирина Евгеньевна Воронина

doi:10.17308/sait/1995-5499/2025/2/78-88

Authors

Nikita A. Ekert Voronezh State University https://orcid.org/0009-0005-8176-8226 (unauthenticated)
Irina E. Voronina Voronezh State University https://orcid.org/0000-0003-0040-5764 (unauthenticated)

DOI:

https://doi.org/10.17308/sait/1995-5499/2025/2/78-88

Keywords:

Large Language Models, LLM, review evaluation, Llama, GPT, Claude, Gemini, GigaChat, Phi3, Gemma, Mistral, DeepSeek, language model comparison, text summarization, meaning extraction, text clustering, Retrieval-Augmented Generation, evaluation criteria generation

Abstract

The article considers the actual problem of setting up input parameters of an algorithm for dividing a text corpus into clusters and issuing names for the identified clusters. The subject of the study is approaches to training large language models. The theoretical significance of the study lies in the providing an assessment of the quality of existing language models and the possibility of their application in solving the problem. The practical significance lies in the software implementation of the algorithm for naming clusters in the automated formation of criteria for evaluating a software product, as well as in conducting a computational experiment to assess the accuracy of the algorithm. The novelty of the result lies in the refinement of the existing algorithm for assessing user feedback.

Author Biographies

Nikita A. Ekert, Voronezh State University

postgraduate student, Department of Software and Information Systems Administration, Voronezh State University
Irina E. Voronina, Voronezh State University

Professor of Software Development and Information Systems Administration Department of Applied Mathematics, Informatics and Mechanics Faculty, PhD, Dr. habil. in Technical Sciences