K-means Clustering In The Credit Rating In The High-tech Enterprises

At information age, the technology competition is getting fiercer and fiercer around the world. In recent decades, the high-tech industry has been developing at a rapid speed, which has obviously become one of the most essential factors for GDP growth. As the high-tech industry is having the profound impact on a nation's technology and economic progress, it has attracted almost all the countries'attention on Earth. However, with the market economy's evolution, the high-tech industry has also exposed some problems, while is although playing the crucial role. Because of the enterprises'growth, the innovation of science and technology, and the demand of transforming into productivity, the high-tech enterprises are faced with financial problems, which are considered to be the key difficult points for these enterprises to succeed.The financial institutions, public policies, and domestic people in China did not supply enough support to the high-tech enterprises during the past years. The reason, apparently, lays in some historical factors, such as high risk, opaque information, low credit security relating to high-tech enterprises. As the matter of fact, owing to the credit rating system has not been established until now, the funding suppliers do not have the chance to clarify the enterprises'credit information, and thus the limited fund could not be distributed to those enterprises which need the financial support urgently. To solve this problem properly, standing on the third party's point, this paper analyzes the high-tech enterprises'credit condition, as well as the change information, expecting to offer some scientific suggestions.The traditional K-means clustering algorithm can handle out large amount of multidimensional data quickly, and can save much time. Whereas due to selecting the initial cluster centers by random, this traditional method sometimes could not gain ideal clustering effects. Considering the shortcoming of this traditional algorithm, using the entropy method, this paper sorts the evaluation objects, makes the difference, and finds the proper initial cluster centers. The results demonstrate that the improved method can acquire better clustering effects, and can have a good application prospects. Hasing rated the enterprises, this paper set up the credit rating transition model of high-tech enterprises. Based on the credit rating transition matrix, this paper analyzes the credit change condition in the past three years, and also predicts what the credit condition would be in the next three years. The results show that the credit risk of high-tech enterprises in the short run would not be so high. But in the long term, the default probability would increase, therefore the financial institutions should pay more attention to those enterprises, controlling credit risk and transferring funds into safer field. Furthermore, the method of calculating credit rating transition matrixes this paper supplied can in some extend pave the way for financial risk management, and would be better for funds supplier.At last, on the basis of research and risk management experience in international financial market, using Markov credit rating transition matrix, this paper takes the CreditMetrics model as example to measure the credit risk in China's financial market.In summary, credit risk management system is widely used and is profound for practical application. Therefore it is the high time to establish the proper method and credit rating system.