会计系学术讲座系列2021年第6讲

主 题：Can Machines Understand Human Decisions?Dissecting Stock Forecasting Skill

主讲人: Sean Cao (佐治亚州立大学副教授)

协调人: 张敏,张博,汝毅

时 间：2021-04-21 10:00

地 点：明德商学楼706会议室

语 言：中英文

讲座摘要:

Human decisions are important but difficult to understand or predict. This paper uses machine learning models, which are adept at capturing nonlinear and complex relations, to analyze analysts’ forecasting decisions and determine their skill. Machine-identified skilled analysts persistently outperform expert-picked star analysts. Machines rely on nonlinear interactions of analyst characteristics, such as past skill and efforts, to identify analyst skill, in contrast with human experts, who lean more on relation-based information such as brokerage size. The puzzle of post-analyst-revision drifts can be explained by our model in that such drifts are concentrated in machine-picked skilled analysts. Our approach also allows the formation of a “smart” analyst consensus that aggregates the forecasts of machine-picked skilled analysts. Investment strategies based on revisions of machine-identified skilled analysts and the smart analyst consensus both generate significant abnormal returns. Overall, we propose an interpretable machine learning framework that can be used to analyze and evaluate human opinions in general settings such as online discussions, political forecasts, and macroeconomic outlooks.

主讲人简介:

Dr. Cao’s research examines the usefulness of firm information and disclosures for investor and corporate decision-making. This research topic is a crucial component of disclosure research in the field of accounting and corporate finance research in the field of finance. A unifying theme in his research is to protect investors’ wealth and increase firms’ operational efficiency and market value using large-scale data analytics of corporate disclosures. Because most corporate accounting and disclosure data are unstructured and high-dimensional, like text and images, emerging technologies such as deep learning and artificial intelligence are crucial for quantifying useful information from this unstructured corporate data using unleashed computational power.

His research has been published in premier journals across finance, accounting, and computer science such as Journal of Financial Economics, Journal of Accounting Research, The Accounting Review, Contemporary Accounting Research, and IEEE Computer. His work has been presented at major research universities, including Stanford University, Cornell University, New York University, University of Virginia, University of Minnesota, and Pennsylvania State University, as well as at national policy-making or advisory agencies such as the SEC and NBER. His research has also received interest and workshop invitations from leading financial and AI firms such as State Street Boston headquarters, Grant Thornton executives, Ant Financial, Baidu, DataYes, and JD.com.

Dr. Cao co-chaired the 2020 GSU-RFS conference on Fintech and Machine Learning. The conference was sponsored by the Review of Financial Studies (dual submission) and Georgia State University. The conference attracts an average of 150 high-quality submissions annually in the area of emerging technologies from preeminent scholars in both finance and accounting.

Dr. Cao’s passion for research carries over into his teaching. He is committed to giving students access to the most cutting-edge learning material to help them achieve career success. He was ranked in the top 10 percent for outstanding teaching university-wide at University of Illinois at Urbana-Champaign. At GSU, he teaches Ph.D. seminars in capital markets research and undergraduate courses in financial accounting with a strong data-analytics component. Dr. Cao has been invited by multiple major research universities to teach short-term doctoral seminars on emerging technologies in finance and accounting.

曹博士的研究聚焦于公司信息披露，特别是公司信息披露对投资者及公司决策的影响。该研究话题同时也是会计学和公司金融领域研究中的重要议题。曹博士的研究成果主要研究观点是，将大数据分析融入到公司信息披露的研究中能够有效保护投资者财富、提升企业的运营效率和公司价值。由于大部分有关公司信息披露的数据，例如文本和图像，具有非结构化和高维度等特征，因此，利用深度学习和人工智能等计算机技术有助于提取和量化公司信息披露中的非结构化信息，而这对于深入挖掘公司的有效信息及其经济后果尤为重要。

曹博士已在Journal of Financial Economics, Journal of Financial and Quantitative Analysis，Journal of Accounting Research, The Accounting Review, Contemporary Accounting Research, and IEEE Computer等金融学、会计学和计算机科学领域的顶级期刊上发表多篇文章。同时，其研究成果也多次在Stanford University, Cornell University, New York University, University of Virginia, University of Minnesota和Pennsylvania State University等著名研究型院校中报告。他的研究也曾在美国证券交易委员会（SEC）和美国国家经济研究局（NBER）等国家政策制定与咨询部门中汇报研究成果。此外，曹博士的研究还吸引了诸多业界公司的兴趣与关注。他的研究曾多次受邀参加State Street Boston headquarters, Grant Thornton executives, Ant Financial, Baidu, DataYes和JD.com等行业领先的金融和人工智能公司的交流研讨。

曹博士是2020年GSU-RFS conference on Fintech and Machine Learning的联席主席。本届会议由Review of Financial Studies (dual submission)与Georgia State University共同资助。作为新兴技术领域的高水平会议，该会议每年平均会吸引150余篇来自金融和会计领域杰出学者的高质量论文。

作为教学与研究并重的学者，曹博士一直致力于在教学中传授最前沿的学科知识，助力学生的职业发展与成就。他曾在University of Illinois at Urbana-Champaign的校级杰出教学评选中位列前10%。在Georgia State University，他将大量的有关数据分析的前沿知识融入到资本市场研讨课、财务会计学等博士和本科课程之中。同时，曹博士还受到多所知名研究型大学的邀请，为相关院校的学生开设短期博士研讨课程，讲授有关新兴技术在金融和会计领域的应用等内容。