BANKRUPTCY FORECAST OF RUSSIAN CONSTRUCTION COMPANIES FOR STABLE DEVELOPMENT OF THE REGIONAL ECONOMY

The issue of forecasting corporate bankruptcy has been relevant in recent years due to economic instability, geopolitical tensions, increased competition, the COVID-19 pandemic, and sanctions pressure. Corporate bankruptcy manifests itself in various forms and scales depending on the industry, region, and general economic situation in the country, and affects the state of the region. The purpose of the study is to develop a model for forecasting the probability of bankruptcy of companies involved in the construction industry. According to the modeling results, the overall predictive ability of the Random Forest model is 82,8%; the Gradient Boosted Trees model is 86,2%; the Tree Ensemble model is 82,8%. Thus, for the purposes of forecasting the bankruptcy of Russian companies in the construction industry, the Gradient Boosted Trees model is most applicable. The developed model for forecasting the probability of bankruptcy of companies can affect the stability, development, and efficiency of the regional economy.