TheFinAI/JF-TE

JF-TE（日本金融术语提取）是一个基准数据集，用于评估从日本专业披露文件中提取和排名嵌套金融术语的层次结构。该数据集包含从10份专业披露文件中提取的202个注释级实例，涵盖了经过标准化的777个独特金融术语的2,412个专家精选术语提及。该任务解决了在混合文本中金融术语边界敏感定位的挑战，其中嵌套复合词和脚本变体外来词（汉字、平假名和片假名）使术语边界和语义范围变得模糊。该数据集提供了金融术语提取的语言学基础评估，反映了日本金融披露文件的复杂性。JF-TE (Japanese Financial Term Extraction) is a benchmark dataset for evaluating hierarchical extraction and ranking of nested financial terminology from Japanese professional disclosures. The dataset consists of 202 note-level instances extracted from 10 professional disclosures, containing 2,412 expert-curated term mentions covering 777 unique finance terms after normalization. The task addresses the challenge of boundary-sensitive grounding of finance terminology in mixed-script text, where nested compounds and script-variant loanwords (kanji, hiragana, and katakana) make term boundaries and semantic scope ambiguous. This dataset provides a linguistically grounded evaluation of financial term extraction that reflects the complexity of real Japanese financial disclosures.