Predicting corporate management performance using AI: Incorporating CEO strategy insights from sustainable management reports

Xiao Wang, Feng Sun, Yong Ki Kim, Hyungjoon Kim, WonHo Song and Yubing Wei

PLOS ONE, 2026, vol. 21, issue 5, 1-23

Abstract: This study proposes an AI-based model to predict corporate management performance by combining financial data with strategic information extracted from CEO messages in sustainability reports. Using a dataset of 1,271 listed companies on Korea’s KOSPI and KOSDAQ markets (2016–2023), we applied eight machine learning and deep learning classifiers: KNN, SVM, GBM, CatBoost, GAN, RNN, LSTM, and Transformer. Financial variables were selected based on prior accounting research, while strategic variables were derived via text mining of CEO messages and categorized using the Sustainable Balanced Scorecard (SBSC) framework. Results show that models incorporating both financial and strategy-based variables outperformed those using financial data alone. Notably, the Transformer model achieved the highest predictive accuracy, followed by LSTM and RNN. These findings provide actionable insights for investors and corporate stakeholders while advancing interdisciplinary research between accounting and AI. Under 5-fold cross-validation, the best-performing hybrid model (Transformer with SBSC features) achieved Accuracy = 0.8467, AUC = 0.8481, and F1 = 0.8572, and adding SBSC strategy indicators improved mean performance across models (ΔAccuracy=+0.0121; ΔAUC=+0.0092; ΔF1=+0.0119).

Date: 2026
References: Add references at CitEc
Citations:

Downloads: (external link)
https://journals.plos.org/plosone/article?id=10.1371/journal.pone.0347140 (text/html)
https://journals.plos.org/plosone/article/file?id= ... 47140&type=printable (application/pdf)

Related works:
This item may be available elsewhere in EconPapers: Search for items with the same title.

Export reference: BibTeX RIS (EndNote, ProCite, RefMan) HTML/Text

Persistent link: https://EconPapers.repec.org/RePEc:plo:pone00:0347140

DOI: 10.1371/journal.pone.0347140

Access Statistics for this article

More articles in PLOS ONE from Public Library of Science
Bibliographic data for series maintained by plosone ().