金融领域事件因果关系抽取数据集

事件抽取是舆情监控和金融领域的重要任务之一。“金融事件”在金融领域是投资分析，资产管理的重要决策参考；事件也是知识图谱的重要组成部分，事件抽取是进行图谱推理、事件分析的必要过程。“事件抽取”的挑战体现在文本的复杂和任务的复杂。文本的复杂体现在事件抽取的输入文本可能是句子、段落或者篇章，不定长度的文本使得限制文本长度的模型无法使用；任务的复杂体现在事件识别的任务包括：事件类型识别，事件要素抽取，事件关系抽取等等。本评测任务的目标是解决篇章级事件元素抽取和事件因果关系抽取这两个核心的知识抽取问题。
本次评测任务的文本语料来自于互联上的公开新闻、报告。在篇章级事件元素抽取任务中，给定篇章级长文本和事件类型，从篇章级文本中识别事件的元素。在事件关系抽取任务中，给定一段描述因果或影响关系的文本，从文本中抽取原因事件的表示和结果事件的表示

Event extraction is one of the core tasks in public opinion monitoring and the financial sector. "Financial events" serve as critical decision-making references for investment analysis and asset management in the financial field; events also constitute an important component of knowledge graphs, and event extraction is a necessary procedure for graph reasoning and event analysis. The challenges of event extraction stem from the complexity of both input texts and the task itself. The complexity of input texts is manifested in that the input materials for event extraction can range from single sentences, paragraphs to full documents, and the variable-length nature of such texts renders models with text length constraints inapplicable. The complexity of the task is reflected in that event recognition encompasses multiple subtasks, including event type identification, event element extraction, event relation extraction and so on. The goal of this evaluation task is to address two core knowledge extraction problems: document-level event element extraction and event causal relation extraction. The text corpora for this evaluation task are sourced from publicly available news articles and reports on the Internet. In the document-level event element extraction subtask, given a long document-level text and an event type, participants are required to identify the elements of the target event from the document. In the event relation extraction subtask, given a text describing causal or influential relationships, participants need to extract the representations of the cause event and the result event from the text.

搜集汇总

背景与挑战

背景概述

该数据集专注于金融领域的事件因果关系抽取任务，旨在从篇章级文本中识别事件元素并抽取原因事件与结果事件之间的因果关系。数据来源于互联网上的公开新闻和报告，适用于舆情监控、投资分析和知识图谱构建等应用场景。数据集包含训练和评估文件，通过事件类型和多个要素（如行业、产品）来表示事件，支持复杂的金融文本分析需求。