Cluster Analysis of Listed Company Solvency Based on the SOM

摘要

Enterprise managers should master the debt risk in order to make the right decision. SOM (Self-Organizing Maps) is a type of artificial neural network trained by unsupervised learning. First, the paper improves the neighborhood function of SOM network. Then, this paper uses the improved SOM network to cluster 24 steel listed companies' solvency. The sort of solvency will offer solution for operators to find insufficiencies for enterprise development.