Font Size: a A A

Systematically Analyzing The Financial Securities Through Quantitative Methods

Posted on:2012-10-20Degree:MasterType:Thesis
Country:ChinaCandidate:S J ZhaoFull Text:PDF
GTID:2189330338494384Subject:Theoretical Physics
Abstract/Summary:PDF Full Text Request
Time series, in the financial industry, can be considered as very important information. The so-called time series is a series of observed data sequencing in chronological order. The observed value sample is taken according to equidistance or non-equidistant time interval. Financial security analysis is often based on historical data or materials, to provide forecast information for future tends to move up and down. For example, stocks, futures opening price, closing price, the highest/lowest price, turnover, volume, etc. can be considered as time series to be analyzed.A series of new indexes is generated including KDJ, MACD, RSI etc, after analyzing, processing and reforming the time series. Forecasting future tendency of securities by these indexes is the frequently-used method at present. Recently in China a type of new quantitative research method is forming. Secondly this method studies the programming trading system. This paper is firstly decomposing and denoising the time series, next working out its relativity and volatility, and finally studying and researching the time series by forecast method (including studying the relativity of bivariate time series end by Copula with multivariate GARCH model). Secondly the paper analyzes memorability of time series by fractal method. Influenced by a lot of nonlinear factors, the financial time series often presents complex morphology and detailed features. Its variation tendency can be analyzed by Hurst index. Moreover, analyzing the time series spectrum can work out the cycle of time series. Linear regression forecasting of time series can be done through the auto-regressive moving average model (ARMA), auto-regressive integrity moving average model(ARIMA), etc.Nonlinear regression forecasting of time series can be done through neural network and SVM (support vector machine) etc. The variables are combined through linear or nonlinear analysis. Then interdependence of variables can be analyzed. As to financial time series, regression analysis is applied to tendency of the sequence and variability of the tendency, which means forecasting the tendency of time series through systematical analysis and study of time series by mathematics, physics, computer intelligent etc.
Keywords/Search Tags:Copula, ARMA, Dynamic correlation, Hurst index, SVM (support vector machine)
PDF Full Text Request
Related items