Human-Reviewed Uzbek Legal Named Entity Recognition Dataset

Abstract

This article describes a human-reviewed Uzbek legal-domain named entity recognition (NER) dataset developed as a reusable resource for low-resource legal NLP. The release contains 12 entity categories: PER, ORG, LOC, DATE, MONEY, POSITION, DOCNO, LAW, COURT, BANK, TIN, and CADASTRE. The dataset is provided in XLSX, CSV, JSON, and JSONL formats and is structured into two complementary layers: a core subset of manually reviewable source-grounded records and an extended augmented subset used to support lower-frequency labels in training-oriented settings. The package also includes supporting documentation, split guidance, a data dictionary, and review-related metadata, including provenance, verification status, and quality flags. Character-level start and end offsets are included where recoverable. The release is intended to facilitate Uzbek legal NER research, resource curation, and transparent reuse under provenance-aware conditions.

More by Bobur Saidov

• UzNER: A Human-Reviewed Benchmark for Uzbek Named Entity Recognition With Gazetteer-Augmented Transformer ModelsJournal of Computer Science · 2026 · Citations: 3
• Entity-Aware Sentiment Analysis for Uzbek Social Media Using XLM-RoBERTa NER and LSTMUral Symposium on Biomedical Engineering, Radioelectronics and Information Technology · 2026
• Software System for Hybrid Sentiment Analysis of Uzbek Texts using Named EntitiesThe Herald of the Siberian State University of Telecommunications and Information Science · 2026
• A Hybrid NER–Sentiment Model for Uzbek Texts: Integrating Lexical, Deep Learning, and Entity-Based ApproachesBig Data and Cognitive Computing · 2026 · Citations: 5