Application Of Vector Generalized Linear Model And Correlation Analysis In Risk Investment

High tech enterprises which are developed by risk investment has become a pillar of the national economy, represented the country’s international competitive, embodied the most advanced science and technology. For the European and American developed countries, the contribution rate of the high-tech enterprises in the national economic growth is 5%-20% in the beginning, slowly rises to 50%, and has reached 60%-80%currently. The push for economic development has far exceeded the traditional capital intensive enterprises. It is really ”science and technology is the first productivity”. The competitiveness of the national economy mainly depends on the level of science and technology and high-tech transformation to form new industry. It’s very important to find the rule of inherence and internal relationship of development of things, then venture capital investors can see clearly of the investment trend as well as the regarding detailed information.Take, for instance, the data of average investment amount of all the funds in the world, the data follows neither the normality assumption nor the homogeneity of variance. We used “analysis of variance estimation of rank” to prove the significant di?erence of average investment by each fund among each year. Assuming the average investment follow the rule of Gamma distribution, the form parameter and scale parameter are quadratic function of year parameter, we can estimate each parameter based on model of generalized linear vector. Finally we forecasted the form parameter of Gamma distribution of 2015 is 1.1959, and scale parameter of Gamma distribution of 2015 is 5816.186. Investment between any intervals can be calculated by probability distribution.For standardized data of average investment per round of investment company and fund and average financing per round of high-tech enterprises, multiple independent test method of Genest 2004 is presented, the results show that the three are not independent of each other. Six Copula, normal Copula, t Copula, Gumbel Copula, Frank Copula, Clayton Copula and Joe copula are chose as candidate models. Gumbel Copula is thought as right one for the risk investment data using the degree of goodness of experience testing method. The parameter is estimated of 1.246. The upper tail and lower correlation coeffcient of fitting Gumbel copula are finally calculated. The results shows that the tail correlation among three is weak. When one of the investment amount is large, the other two are not necessarily great.A variety of correlation coeffcients are applied to find the dependence strength of the investment company, investment fund and high-tech enterprises per round. Two rank correlation coeffcients such as Spearman ’s’ s Rho and Kendall tau, improved spearman’s Rho coeffcient, mutual information index, Distance correlation, HHG distance(Heller Heller Gorfine Distance) and the maximum information coeffcient(MIC)get the same conclusion. MIC is published in 2011 ”Science”. high-tech enterprises financing is highly correlated with fund’s investment and poorly with company’s investment. Investment between company and fund seems random. High tech enterprise financing is mainly through the risk fund,and less directly from VC firm. For Vc firm,direct investment and the fund investment has no fixed pattern.A comprehensive evaluation method of high growth company of science and technology is established. The method is appreciated for growing or mature IDG company,especially for our medical group. Based on the above analysis, 12 kinds of net profit rate are forecasted, and 65000000 yuan net profit is get. In the neutral position, the three methods, DCF method, net present value method and internal rate of return method are used for analysis of investment return. We modify the practice case of three modes.Through the analysis of Shuangcheng Medicines Co’s case, the method is demonstrated. A year later, the financial data of Shuangcheng medicine in 2010 shows that net profit reached 66000000 yuan. This is consistent with our prediction data of 65000000 yuan, and sales are similar.