Time Sequence Analysis Of Financial Time Series And The Corresponding Research Of Risk Measurement

Risk measure of financial time series has been a hot topic in financial Studies. Since Markowitz proposed mean- variance to describe financial risks, financial risk measurement research has been developed. In recent years, new indicators have been put up to describe the risks of financial time series, such as Va R, CVa R,etc.For the indicators to describe time series risks, whether the variance, Va R or CVa R, variance is always used to characterize the volatility and uncertainty of the time series, and the risk values are independent of the order of the time series. But this article has demonstrated that the risks are associated with the sort order of the financial time series, therefore characterizing the volatility and uncertainty of time series with variance is limited. Domestic and foreign scholars studied little on the timing characteristics of the time series, and there is no systematic study describing the timing characteristics. For this reason, the article studied timing characteristics of the time series the risk of financial time series and demonstrated the volatility and uncertainty of different demonstrated time series. About How to introduce timing characteristics to the risk value calculation of financial time series, this paper used two methods. First, we Designed timing indicators time variance TV2 to describe the uncertainty and volatility of the time series, and then with it we calculated the new risk values; Second, we established the ARMA-GARCH model to get the conditional variance. In this paper, CVa R value is used to characterize the risks of financial time series, and with the two proposed methods we optimized the calculation of CVa R value.For calculation method of CVa R value, we choose the Monte Carlo simulation method to calculate CVa R. Because the scholars home and abroad generally choose standard normal distribution to describe the market factors model of Monte Carlo simulation, which cannot reflect the feature of aiguille and fat tail. This paper apply GED distribution to Monte Carlo simulation method, and combine Monte Carlo with the two time sequence adjusted methods to established TV-MC model and ARMA-GARCH-MC model.To verify the two models established in this paper, we selected closing price index of Shanghai Composite Index from February 4, 2010 to March 5, 2015 for the empirical study and compared the results with the traditional model which is not adjusted by time sequence. The results showed that the two models are better than the traditional model which is not adjusted by time sequence. Because the two models have different limitations, this article combined the two models to establish ARMA-TV-GARCH-MC model, and got more satisfactory results.