metadata

license: apache-2.0
task_categories:
  - text-classification
language:
  - zh

FinTruthQA: A Benchmark Dataset for Evaluating the Quality of Financial Information Disclosure

About

FinTruthQA is a comprehensive benchmark designed for evaluating the quality of financial information disclosure in Q&A data, particularly within the context of Chinese stock exchanges' investor interactive platforms. These platforms enable listed companies to respond to investor inquiries in an online Q&A format. However, it is common for listed firms to respond to questions with limited or no substantive information, and automatically evaluating the quality of financial information disclosure on large amounts of Q&A pairs is challenging. FinTruthQA comprises 6,000 real-world financial Q&A entries and each Q&A was manually annotated based on four conceptual dimensions of accounting: question identification, question relevance, answer readability, and answer relevance. By establishing this benchmark, we provide a robust foundation for the automatic evaluation of information disclosure, significantly enhancing the transparency and quality of financial reporting. FinTruthQA can be used by auditors, regulators, and financial analysts for real-time monitoring and data-driven decision-making, as well as by researchers for advanced studies in accounting and finance, ultimately fostering greater trust and efficiency in the financial markets.

For more details, please refer to our paper.

Authors: Ziyue Xu, Peilin Zhou, Xinyu Shi, Jiageng Wu, Yikang Jiang, Bin Ke, Jie Yang

Dataset

We collected Q&A entries from the interactive platforms for communication between investors and listed companies established by the Shanghai Stock Exchange(SSE) and the Shenzhen Stock Exchange(SZSE). The data from the SSE covers the timeframe from January 4, 2016, to December 31, 2021, while the data from the SZSE spans from September 1, 2021, to May 31, 2022. Each Q&A entry was annotated based on four key information disclosure quality evaluation criteria: question identification, question relevance, answer relevance, and answer readability.

Data Statistics

The figure below shows the distribution of questions and answers in FinTruthQA(in characters).

Cite

If you use FinTruthQA in your paper, please cite our paper:

  @misc{xu2024fintruthqa,
  title={FinTruthQA: A Benchmark Dataset for Evaluating the Quality of Financial Information Disclosure}, 
  author={Ziyue Xu and Peilin Zhou and Xinyu Shi and Jiageng Wu and Yikang Jiang and Bin Ke and Jie Yang},
  year={2024},
  eprint={2406.12009},
  archivePrefix={arXiv},
  primaryClass={cs.CL},
  url={https://arxiv.org/abs/2406.12009}, 
  }

Contact

If you have any questions about our dataset or benchmark results, please feel free to contact us!

(Ziyue Xu, [email protected]; Jie Yang, [email protected])