Study On VaR Of Finacial Time Series And Construction Of Web Based Data Visulization System

The abundant financial derivatives, coupled with the emerging Internet Finance, makeindividuals and enterprises face more investment choices. Risks and benefits are always tiedtogether and restricted each other. How to obtain the most benefit with minimum risk underthe condition of limited capital is a long-term problem that is in front of academia, financeand investors. It is also the ultimate goal and research content of Portfolio Theory. Financialtime series analysis is an interdiscipline that involves in mathematical statistics, economics,finance and computer science. Naturally, we can improve the portfolio model by studying thecharacteristics and risk models of financial time series.To begin with, research background and significance of this subject were introduced inbrief, and the related literature and techniques at home and abroad were reviewed andsummarized. Combined with Generalized Autoregressive Conditional heteroscedasticity(GARCH) model and Markov Chain Monte Carlo (MCMC) method, an improved VaR (Valueat Risk, VaR) method was propoed, by which method, an empirical study and verification wascarried out using the CSI300Index data.Next, double parameters q-gaussian probability density function expression is derivedfrom the linear differential equation, then the graphical features and parameters estimationmethod was studied. Aiming at the stock yield data distribution characteristics, doubleparameters q-gaussian can be applied in the classical portfolio model. Real stock empiricalresearch shows that greater gains can be achieved with the proportion calculated byq-gaussian based portfolio model than that without q-gaussian.Then, to describe the dependent constructure between assets yield sequences and to solvethe heavy-tailed distribution phenomenon, different copulas、extreme value theory and q-gaussian distribution are introduced to construct an integrated framework to model the assetsyield sequences and measure portfolio risk and gains. Eight economy indexes which areclosely related to China’s economy in recent years were selected to verify the effectiveness ofthe proposed integrated framework. The empirical results demonstrate that the framework caneffectively reduce the investment risk, and most importantly it can lead the most benefitscompared with others like it.Eventually, a web based financial data visualization system has been constructed basedon the excellent performance of R language in data analysis, the characteristics of InternetFinance, as well as the latest data visualization technology such as RShiny, FastRWeb andECharts.