Non-stationary Long Memory Signal Modeling And Forecasting

This paper revolves around chaotic signal theory,we introduced the concept to study the complex system signal having long-term memory.Based on this concept we found that the velocity of two time interval and the corresponding force since the co-variance is equal to zero,thus can get complicated system signal effective about two times the interval velocity of conditional probability distribution function.Further,by this study we found that the velocity of the continuous time interval of this can be an effective about two times the interval velocity of this export,on the basis of some basic properties of it we could deal with complex system signal statistical forecasting.Based on the method,we use the CSI 300 stock index futures time series signal conditional probability distribution function of the time series model to forecast the future,and operator inspection within the time limit.We collected from April 2010 to July 2016 the details of the CSI 300 stock index futures time series signal as the research object.Through displacing to the price time series signal of Shanghai and Shenzhen 300 index futures(velocity)-time curve,the anomalous diffusion,the correlation functions and yields Tsallis-q-Gauss distribution four aspects carries on the analysis,research the anomalous statistical properties.Due to the asymmetry of the asking price of conditional probability,we need to use symmetry of conditional probability of asymmetric conditional probability,then through q for symmetric gaussian distribution we described in detail the conditional probability of fitting,and then obtained the desires of conditional probability,the image and the actual data fitting very well.Finally through calculating the CSI 300 stock index futures time series signal,from data fitting,in any case this can be in only six parameters to parse said.This analysis can also be used to calculate the conditions of the conditional expectation and velocity variance.It is interesting to note that the data is displayed in twoadjacent time interval,the velocity of the conditional expectation is monotonous,and conditional variances tends to dull.In addition,modeling and forecasting precision of the method,also the principle of simple and intuitive,calculation results are moderating,the extrapolation model generalization ability is stronger.The above content is for this paper research content.It is worth noting that this general expression,not only suitable for the current stock futures system,the other long-term sexual signals of complex systems.