Work

Classification of Uzbek Banking Texts by Document Types

Year: 2026

Type: conference-paper

Abstract: The article describes a method for automating text classification by document types using logistic regression and N-gram-based text vectorization. It should be noted that the current study focuses on developing an algorithm that can process both simple and complex text data. Including texts that have different thematic contexts. Moreover, for the most objective assessment of the developed tool, three data sets were prepared, each with its own degree of complexity. As a result of testing, it was found that the algorithm showed fairly high accuracy on texts with a clear structure, in particular, the best value of metrics was achieved when using early stopping, as well as hyperparameter optimization. In addition, authors organized detailed comparative analysis of existing alternative works, which can help to see current actual problems of the considered topic. In the future, the authors plan to implement the proposed approach using neural network technologies, which will significantly expand the capabilities of this algorithm.

Source: International Siberian workshops and tutorials on electron devices and materials

DOI: 10.1109/edm69524.2026.11631992

Authors: Nilufar Abdurakhmonova, Sadokat Nurmanova, Mushtariy Kozimjonova, Salomat Otamurodova, Mokhinur Toshpulatova, Alisher Abidjanov

Institutions: National University of Uzbekistan, Andijan State University, Tashkent State Agrarian University, Tashkent State University of Economics, Cyber University

Language: English

Topic: Text and Document Classification Technologies

Subfield: Artificial Intelligence

Field: Computer Science

Domain: Physical Sciences

Open Access status: hybrid

OpenAlex ID: W7202313880

FWCI:

Cites: 19

Cited by:

Related to: 10