ESDC: An Earth Science Data Corpus to Support Literature Information Extraction

Large amounts of original research data related to earth system science have increased rapidly over the past ten years, making it easy to help researchers understand the human-earth system from different fields. A substantial amount of these earth science data products is published as open source by geographers in authoritative journals. Extracting information from the literature of these authoritative journals holds significant potential in constructing a domain knowledge base. However, this potential remains largely unfulfilled in geoscience. This is because there is almost no publicly available related corpus and baselines, which is one of the biggest obstacles. Therefore, we present the Earth Science Data Corpus (ESDC), an academic text corpus of 500 abstracts from the authoritative journal. To the best of our knowledge, it is the first corpus of earth science data to provide such detail, which provides a professional training dataset for knowledge extraction and construction of domain-specific knowledge graphs from massive literature. In the production process of ESDC, we considered the contextual features of spatiotemporal entities and the linguistic characteristics of academic literature and formulated annotation guidelines and procedures tailored for earth science data to ensure the reliability of the ESDC.

近十年来，地球系统科学领域的原始研究数据规模快速扩张，为研究者跨领域认知人地系统提供了有利条件。该类地球科学数据产品中，有相当一部分由地理学者在权威期刊中以开源形式刊发。从上述权威期刊文献中提取信息，对于构建领域知识库具有重要价值，但这一潜力在地球科学领域尚未得到充分发掘。究其原因，目前几乎没有公开可用的相关语料库与基准数据集，这是制约该领域发展的最大障碍之一。为此，我们构建了地球科学数据语料库（Earth Science Data Corpus, ESDC）——这是一个包含500篇权威期刊摘要的学术文本语料库。据我们所知，这是首个针对地球科学数据、提供此类详尽信息的语料库，可为从海量文献中开展知识抽取与领域专用知识图谱构建提供专业训练数据集。在ESDC的构建过程中，我们充分考量了时空实体的上下文特征与学术文献的语言特性，并针对地球科学数据制定了专属的标注规范与流程，以确保语料库的可靠性。

搜集汇总

背景与挑战

背景概述

ESDC是一个地球科学数据语料库，包含500篇权威期刊摘要，旨在支持文献信息提取和领域知识图谱构建。该数据集首次提供了详细的地球科学数据标注，考虑了时空实体的上下文特征和学术文献的语言特点，确保了数据的可靠性。