Research of Financial Early-warning Model for the Listed Electric Power Companies on Evolutionary Support Vector Machines Based on Genetic Algorithms

Abstract

The development of the electric power enterprise concerns the national economic lifeline. In this paper, the Support Vector Machines (SVMs) early-warning model which is based on Genetic Algorithm (GA) optimization is established, with GA’s ameliorating SVMs. Using the penalty parameters and the kernel parameters of the process of GA’s optimizing SVMs, this paper gives full play to the global searching ability of GA and overcomes the problems generated from the selection of the SVMs model parameters. As a result, it is possible to initiate the financial risk analysis of the electric power enterprises and enable them to take timely measures to deal with issues that have emerged during the process of their development. It is displayed in the instance verification results of the listed companies in the electric power industry that SVMs which are based on GA optimization can predict the financial risks of the listed companies in the electric power industry accurately and effectively.