Research On The Construction And Application Of Time Series Network Based On Fuzzy Information Granule

Time series refers to the sequence of values of the same statistical index in the order of their occurrence.The main purpose of analyzing time series is to predict the size or trend of data in a certain period of time according to the known historical data.According to the time of observation,the time in a time series can be a year,quarter,month,or any other form of time.Time series exist widely in finance,science and other fields.The researchers used statistical methods to analyze the time series.The long-term fluctuation of time series is generally composed of a large number of short-term behaviors with variant dynamical characteristics,where kinds of fluctuation patterns in different periods change mutually.By studying and analyzing the internal fluctuation mechanism of observed data,relevant models are established to implement the prediction.In the real world,the obtained data in most cases are inaccurate and vague,so it would be took a lot of time,manpower and material resources to build a time series prediction model based on accurate data.It is of great importance to establish a model which based on the fuzzy time series to predict the future and make correct decisions,whether in the field of research or in production and life.Therefore,in this paper,we propose a novel method to construct fuzzy information granular in polar coordinates and achieve the prediction of long-term time series on the basis of the short-term fluctuation patterns.In the first group of experiments,a new method of constructing fuzzy information granules is proposed,which is to construct fan-shaped fuzzy information granules(PFIG)in polar coordinates,In addition,sliding time windows of different lengths are set for experiments to analyze the transmission characteristics of time series.In the second group of experiments,the short-term fluctuation patterns of time series are studied based on the fan-shaped fuzzy information granules.On this basis,the long-term time series are predicted at the granularity level.The validity of the proposed model is verified on different types of data set.The third group of experiments studied the dynamics of the network and applied it.The contributions and innovations of this paper mainly include the following aspects:(1)The fan-shaped fuzzy information granule is first proposed in order to depict trend information and its changes for each segment of time series.Time series are divided into segments by means of the sliding time window,and fuzzy information granules are defined based on the regression models to indicate the fluctuation pattern of each segment.Compared with the existing models,the fan-shaped fuzzyinformation granule can represent the trend and fluctuation range of data.Then,the fuzzy relationship between information granules is mined and the fuzzy reasoning system is constructed.Prediction is executed at the granularity level to achieve the goal of long-term prediction.(2)Based on the fan-shaped fuzzy information granules,a dynamical transmission networks can be established,by which the transmission mechanism of fluctuation patterns of time series is revealed.The constructed networks are introduced to capture the transmission characteristics of fuzzy information granules.The main advantage of this method is to observe the fluctuation characteristics of time series in a global way,that is,the fluctuation law of data over time.The results show that only a few types of fuzzy information granules and fuzzy relation groups play a key role in the fluctuation mechanism,which always have specific targets.Hence,according to the distribution of the transmission probability,a prediction scheme on the level of information granules can be established.(3)Based on the paroxysmal dynamics analysis of time series network with fuzzy information granules,analyzing clustering coefficients and counting the degree of nodes in the network.It is found that the network conforms to power law distribution.This paper studies the impact of paroxysmal and paroxysmal on information transmission.By introducing an empty model,setting different infection thresholds and experimenting on different data sets,it draws a conclusion that paroxysmal influences information transmission in the network.