An Approach of Enterprise Financial Crisis Pre-warning Based on Multi-class Support Vector Machine

摘要

An approach based on multi-class support vector machine (SVM) is put forward for pre-warning enterprise financial crisis. On the basis of financial crisis pre-warning index system, the actual situations of enterprises’ financial crisis are used as learning samples to train and test the learning machine through cross-validation and one to one vote methods and all the optimal hyper plane functions of every two classes can be obtained. An example testifies the efficiency and feasibility of the model, which provides a new way for dynamic pre-warning of enterprise financial crisis.