Purpose

The purpose of this paper is to give a comprehensive review and synthesis of automated textual analysis of corporate disclosure to show how the accuracy of disclosure tone has been incremented with the evolution of developed automated methods that have been used to calculate tone in prior studies.

Design/methodology/approach

This study have conducted the survey on “automated textual analysis of corporate disclosure and its impact” by searching at Google Scholar and Scopus research database after the year 2000 to prepare the list of papers. After classifying the prior literature into a dictionary-based and machine learning-based approach, this study have again sub-classified those papers according to two other dimensions, namely, information sources of disclosure and the impact of tone on the market.

Findings

This study found literature on how value relevance of tone is varied with the use of different automated methods and using different information sources. This study also found literature on the impact of such tone on market. These are contributing to help investor’s decision-making and earnings and returns prediction by researchers. The literature survey shows that the research gap lies in the development of methodologies toward the calculation of tone more accurately. This study also mention how different information sources and methodologies can influence the change in disclosure tone for the same firm, which, in turn, may change market performance. The research gap also lies in finding the determinants of disclosure tone with large scale data.

Originality/value

After reviewing some papers based on automated textual analysis of corporate disclosure, this study shows how the accuracy of the result is incrementing according to the evolution of automated methodology. Apart from the methodological research gaps, this study also identify some other research gaps related to determinants (corporate governance, firm-level, macroeconomic factors, etc.) and transparency or credibility of disclosure which could stimulate new research agendas in the areas of automated textual analysis of corporate disclosure.

中文翻译：

---

根据不同自动化方法的演变对公司披露文本分析的回顾

目的

本文的目的是对公司披露的自动文本分析进行全面的综述和综合，以显示随着以前研究中用来计算基调的自动方法的发展，披露基调的准确性是如何提高的。

设计/方法/方法

这项研究通过在2000年之后在Google Scholar和Scopus研究数据库中进行搜索来准备“论文披露的自动文本分析及其影响”调查，以编制论文清单。在将现有文献分类为基于字典和基于机器学习的方法之后，本研究再次根据其他两个维度对这些论文进行了分类，即公开的信息来源和语气对市场的影响。

发现

这项研究发现了有关使用不同的自动方法和使用不同的信息源如何改变音调的价值相关性的文献。这项研究还发现了有关这种语气对市场影响的文献。这些都有助于研究人员进行投资者的决策和收益与回报预测。文献调查表明，研究差距在于方法学的发展，以便更准确地计算音调。这项研究还提到了不同的信息来源和方法如何影响同一家公司的披露基调，进而可能改变市场表现。研究差距还在于找到具有大规模数据的公开基调的决定因素。

创意/价值

在回顾了基于公司披露的自动文本分析的一些论文之后，本研究表明了结果的准确性如何随着自动方法论的发展而增加。除了方法学研究方面的空白外，本研究还确定了其他与决定因素有关的研究空白（公司治理，公司水平，宏观经济因素等）以及披露的透明度或公信力，这可能会刺激自动化文本研究领域的新研究议程。公司披露分析。