Securities Time Series Information In The Singular Point Of Research And Modeling

With the development of economy and the conversion of people's investment consciousness, stock investment becomes an important part of people's life in modern times. Stock market is a high-risk and high-interest domain, therefore through researches into its internal disciplinarian, effective analytic methods and tools are being looked for more interest with lower, risk. Therefore the study of disciplinarian in financial time series has great theoretical significant and applicable value.Modeling methods are generally used in fitting security time series. This paper fits a period of security time series with both neural network and time series model. The results indicate the methods are effective in clear tendency, but singularities changing the trend of time series can not be fitted well. The singularities are turning point, including sharp rise points, slump points, extreme points,.etc. The analysis conclusions as follows: First, change rate with relative errors is nonlinear correlation, the higher change rate, the lager relative errors. The errors of singularity have poor performance simulated by white noise or random errors in traditional way. Second, introduce the theory of singularity initiated by security information, describing fluctuation singularity with effect intensity of security information. Third, present a new model base on the theory of security information singularity, introducing information singularity in security time series.This paper does further research of security information and the singularity by the data analysis of Shanghai stock index and Changhe stock, and finally fitting data with the model of security information singularity. The new model combines traditional methods with security information singularity effectively and improves the fitting accuracy up to 50%, and then the singularity could be theoretically explained. It outperforms traditional models on security time series containing many uncertain factors. In summary, it is higher in practicality and efficiency, which becomes a new research direction in security time series.