FinGen

FinGen数据集由日本的人工智能研究中心开发，专注于金融领域的论点生成。该数据集包含三种任务：Evidence2Claim、Chart2Argument和News2Argument，分别涉及基于证据生成未来操作声明、基于价格图表生成投资建议摘要和基于新闻文章生成未来情景及投资建议。数据集大小总计43,924条，涵盖英语和中文文本及图表数据。创建过程涉及从不同来源收集数据，如财务会议记录、投资博客和专业分析报告。FinGen旨在通过模拟金融领域中的未来预测和决策过程，解决金融分析和投资决策中的自动化问题。

The FinGen dataset, developed by the Artificial Intelligence Research Center of Japan, focuses on argument generation in the financial domain. It includes three tasks: Evidence2Claim, Chart2Argument, and News2Argument, which respectively cover generating future operational statements based on evidence, generating investment advice summaries based on price charts, and generating future scenarios and investment advice based on news articles. The dataset has a total of 43,924 entries, covering English and Chinese texts as well as chart data. Its creation involves collecting data from diverse sources such as financial meeting minutes, investment blogs, and professional analysis reports. FinGen aims to address the automation challenges in financial analysis and investment decision-making by simulating the future prediction and decision-making processes within the financial sector.