Construction of Financial Crisis Early Warning Model of State-Owned Enterprises Based on Sparse Clustering Algorithm
Abstract
The fierce rivalry, intricate and ever-changing market dynamics, as well as the inherent difficulties, offer difficult conditions for the growth of state-owned firms in India. The financial crisis appears as a significant element determining these businesses' survival chances. This problem profoundly jeopardizes the long-term functioning and advancement of businesses. Failure to forecast and understand it might have serious consequences for both businesses and society. In recent years, government authorities have regularly issued directives and actions, highlighting the necessity for state-owned firms to strengthen their financial crisis management and control. This study proposes a sparse clustering approach to construct the initial equivalence class based on the restricted similarity of objects and the equivalence connection concept. This initial equivalence connection is subsequently refined using the similarity of equivalence relation, resulting in a more logical grouping outcome. This is the foundation of the state-owned enterprises' financial crisis early warning model. The model achieved an amazing forecast accuracy rate of 84.48% for ST firms that were marked for special treatment three years before, based on research and empirical testing on ST companies in the domestic securities market. This demonstrates the model's strong predicting ability and practical utility.