In recent years, increasing economic volatility, the complexity of financial structures, and the growing uncertainty in business environments have made the early detection of corporate bankruptcy signals a crucial issue in finance and risk management. This study aims to design an early warning system for corporate bankruptcy using big data analytics and by simultaneously leveraging both financial and non-financial data sources. In this regard, the integration of firms’ financial information, behavioral data, and large-scale textual data is employed to identify hidden patterns associated with bankruptcy risk.
The research methodology is based on big data analytics and the application of machine learning algorithms within classification frameworks capable of distinguishing firms into three states: healthy, financially distressed, and bankrupt. Furthermore, data preprocessing, cleaning, and feature selection techniques are applied to enhance model accuracy and efficiency. This approach enables the simultaneous analysis of both structured and unstructured data.
Previous studies indicate that machine learning-based models outperform traditional statistical methods in predicting financial distress. Moreover, the use of big data can facilitate the early detection of warning signals and improve the accuracy of predictive systems. Overall, the findings suggest that integrating financial data with behavioral and textual big data can lead to the development of a more accurate, intelligent, and efficient early warning system for financial risk management.