Transfer Learning and Textual Analysis of Accounting Disclosures: Applying Big Data Methods to Small(er) Data Sets

39 Pages Posted: 20 Apr 2020

See all articles by Federico Siano

Federico Siano

Boston University Questrom School of Business

Peter D. Wysocki

Boston University Questrom School of Business

Date Written: March 20, 2020

Abstract

We introduce and apply machine transfer learning methods to analyze accounting disclosures. We use the examples of the new BERT language model and sentiment analysis of quarterly earnings disclosures to demonstrate the key transfer learning concepts of:

(i) pre-training on generic “Big Data”,

(ii) fine-tuning on small accounting data-sets, and

(iii) using a language model that captures context rather than stand-alone words.

Overall, we show that this new approach is easy to implement, uses widely-available and low-cost computing resources, and has superior performance relative to existing textual analysis tools in accounting. We conclude with suggestions for opportunities to apply transfer learning to address important accounting research questions.

Keywords: Disclosure, Earnings Announcement, Machine Learning, Natural Language Processing, Neural Network, Textual Analysis, Transfer Learning

JEL Classification: G31, G32, M21, M41

Suggested Citation

Siano, Federico and Wysocki, Peter D., Transfer Learning and Textual Analysis of Accounting Disclosures: Applying Big Data Methods to Small(er) Data Sets (March 20, 2020). Available at SSRN: https://ssrn.com/abstract=3560355 or http://dx.doi.org/10.2139/ssrn.3560355

Federico Siano (Contact Author)

Boston University Questrom School of Business ( email )

595 Commonwealth Avenue
Boston, MA 02215
United States

Peter D. Wysocki

Boston University Questrom School of Business ( email )

595 Commonwealth Avenue
Boston, MA 02215
United States

Here is the Coronavirus
related research on SSRN

Paper statistics

Downloads
374
Abstract Views
851
rank
87,756
PlumX Metrics