Named Entity Recognition for Uzbek Economic and Financial Text

Abstract

This paper proposes a combined approach to recognizing named entities in Uzbek economic and financial texts. The proposed system includes several stages: script normalization (Cyrillic → Latin), additional dialect standardization (converting Oghuz/Khorezm words to their standard Uzbek equivalents), and then application of a neural model based on mBERT. A corpus of over 3,000 annotated sentences from regulatory acts, news articles, and financial reports was prepared for training. Furthermore, it should be noted that the annotation was performed using the BIOES scheme, taking into account the morphological features of the Uzbek language. The results demonstrate that combining rulebased modules with a neural model improves the recall and accuracy of entity extraction compared to purely neural solutions. The proposed approach can be used to automate the processing of Uzbek texts-from information retrieval to document digitization-and demonstrates the potential for adapting NER systems to languages with limited resources and rich dialectal variation.

More by Nilufar Abdurakhmonova

• A model of a universal linguistic knowledge graph for Turkic languagesArctic XXI century · 2026
• Machine Learning-Based Topic Classification of Economic and Financial Texts in UzbekInternational Conference on Computer Science and Engineering · 2026
• Building a Phraseological Ontology of Uzbek Idioms and Evaluating Similarity Measures for Their Detection in TextInternational Conference on Computer Science and Engineering · 2026
• Developing Morphological and Syntactic Analysis Models for Karakalpak as a Low-Resource Turkic LanguageInternational Conference on Computer Science and Engineering · 2026