Construction of Financial Crisis Early Warning Model of State-Owned Enterprises Based on Sparse Clustering Algorithm

The fierce competition, complex and changeable market environment and its own inherent particularity make the development of state-owned enterprises face unfavorable situations. The financial crisis is an important factor that affects the survival crisis of state-owned enterprises in China. Financial crisis fundamentally threatens enterprises' sustainable operation and development, and if it is not predicted and analyzed, it will probably bring serious impact to enterprises and society. In recent years, government departments have frequently issued documents and measures to require enterprises, especially state-owned enterprises, to strengthen the management and control of financial crises. The method adopted in this paper is a sparse clustering algorithm, which forms the initial equivalence class based on the scant similarity between objects and the principle of equivalence relation, and corrects the initial equivalence relation through the similarity of equivalence relation, so that the clustering result is more reasonable, thus constructing the financial crisis early warning model of state-owned enterprises. Through the research and empirical test of ST company in the domestic securities market, the prediction accuracy rate of ST company was 84.48% three years ago when it was declared special treatment. This shows that the model has good forecasting ability and practical significance