Ask the Accounting Community

Ahmad H. Juma'h, Ph.D., CPA, CMA

Professor of Accounting

Elham Khorasani Buxton, Ph.D.

Associate Professor of Computer Science

Shashank Mani Tripathi, MS

Computer Science

University of Illinois Springfield

Springfield, Illinois, USA

Abstract

This study introduces an advanced impact materiality disclosure framework for breached companies by integrating large language models (LLMs) and ensemble learning techniques. Using Management Discussion & Analysis (MD&A) sections from 10-K filings, we develop predictive classifiers based solely on publicly available financial data from firms reporting data breach incidents in their financial statements. Unlike traditional regression-based models, our approach enhances predictive accuracy by leveraging natural language processing (NLP) to extract materiality reporting cues related to data breach features. Further we use predictive analysis and machine learning (ML) to predict impact on performance. This hybrid methodology combines structured financial data with unstructured textual insights, offering a scalable, cost-effective solution for identifying financial impact materiality reporting. Our findings contribute to academia and industry by advancing predictive analytics for materiality disclosures, improving risk assessment, and strengthening financial transparency in publicly traded companies.