Omnipresent AI and Big data for financial early warning: integrating financial indicators and text sentiment analysis in Chinese real estate

This study proposes a comprehensive framework which integrates big data analytics and omnipresent artificial intelligence (AI) to predict financial distress in Chinese real estate enterprises. The research examines 118 listed firms through six machine learning algorithms with explainable AI to assess both financial indicators and text-linguistic features. The empirical results demonstrate the LightGBM model's superior performance, while textual analysis—particularly Management Discussion and Analysis (MD&A) sentiment—enhances predictive accuracy. Critical determinants include net assets growth rate, accounts receivable ratio, and profitability metrics. Financially distressed enterprises show heightened optimistic lexical patterns in corporate disclosures, with implications for stakeholder decisions.