【SPEECH】Text Mining in Accounting

On May 16th, the Accounting II course from IBBA, taught by Professor Wil Martens welcomed Dr. Sheng-Feng Hsieh, an Assistant Professor from the Department of Accounting, at the College of Management, National Taiwan University. Dr. Hsieh, who teaches Accounting Information System for both undergraduate and graduate programs, began the lecture by sharing his fascinating academic and career journey.

Dr. Hsieh then explained how text mining in accounting can be performed using various sources such as 10-K filings, press releases, conference call transcripts, analyst reports, auditor’s reports, ESG reports, and social media. Text mining aims to extract data and identify meaningful patterns to evaluate the quality of information.

He introduced the "dictionary-based approach" in text mining, where pre-defined dictionaries of positive and negative words are used to analyze texts. While this method is straightforward, it sometimes overlooks contextual nuances. Therefore, regularly updating the dictionaries is essential to maintain accuracy and prevent misinterpretation.

Dr. Hsieh was asked if all companies issue XML financial statements, to which he replied that companies produce "human-readable" financial reports for public accessibility. Another question addressed whether companies use text mining to adjust reports. Dr. Hsieh mentioned that while direct evidence is rare, indirect evidence can be observed through market reactions and textual analysis.

This guest lecture had a positive impact on IBBA students, enhancing their analytical skills for their final projects. The insights gained from text mining can significantly improve the quality of future accounting literature, as demonstrated by the students' enthusiastic participation and inquiries during the session.

Written by Priscilla Diani Kumaradewi

(Provided by IBBA)