Medya Earnings Call Dataset

该数据集由Medya等人整理，涵盖了2010年至2019年期间的财报电话会议数据，分为五个行业板块：金融、医疗、基础材料、服务和科技。该数据集主要用于分析财报电话会议对股票价格趋势的即时财务影响，特别关注基于每日价值的变化标签。研究任务是基于财报电话会议进行二元趋势标签预测。

This dataset was curated by Medya et al., covering earnings conference call data spanning 2010 to 2019 and divided into five industry sectors: finance, healthcare, basic materials, services, and technology. It is primarily employed to analyze the immediate financial impact of earnings conference calls on stock price trends, with a particular focus on daily value-based change labels. The core research task supported by this dataset is binary trend label prediction based on earnings conference call data.

搜集汇总

背景与挑战

背景概述

该数据集是用于金融长文档分类和趋势预测的收益电话会议记录集合，包含CSV格式的原始数据、AMR格式的语义图数据以及预训练模型权重。数据集支持'FLAG'方法的研究，该方法通过基于AMR的图神经网络进行金融文档分析。