UD-KSL Treebank v1.3
收藏资源简介:
UD-KSL Treebank v1.3是一个针对第二语言(L2)韩语的半自动化框架,用于从XPOS序列中识别形态句法结构,并将其与相应的UPOS类别进行对齐。数据集扩展了现有的L2-Korean语料库,包括从议论文中标注的2998个新句子。数据集旨在解决形态句法标注和依赖解析准确性的问题,特别是在标注数据有限的情况下。该框架利用了韩国eojeol的结构,解释了不同的词素如何组合形成特定的形态句法类别。数据集通过在标注层之间提高一致性,增强了形态句法标注和依赖解析的准确性。
UD-KSL Treebank v1.3 is a semi-automated framework for second language (L2) Korean. It is designed to identify morphosyntactic structures from XPOS sequences and align them with corresponding UPOS tags. This dataset expands the existing L2-Korean corpus by including 2998 newly annotated sentences extracted from argumentative essays. The dataset aims to address the issue of low accuracy in morphosyntactic annotation and dependency parsing, especially when labeled data is limited. The framework leverages the structural characteristics of Korean eojeol, explaining how different morphemes combine to form specific morphosyntactic categories. By improving consistency across annotation layers, the dataset enhances the accuracy of morphosyntactic annotation and dependency parsing.

- 1UD-KSL Treebank v1.3: A semi-automated framework for aligning XPOS-extracted units with UPOS tagsUniversity of Oregon, University of Illinois Chicago, Konkuk University, Chung-Ang University, Yale University · 2025年



