The Reseach On Dependence Of Interent Finance Concept Stocks Based On Time-varying Copula Model With Wavelet Analysis

In recent years,China's Internet financial concept stock market continues to strengthen.But compared with the mature developed countries,the proportion of risk is high,and the synchronization of stock price changes is still very high which is also a typical strong noise market.With frequent policy changes,the change trend of the stock market is very unsmooth.It is necessary to introduce wavelet analysis to detect trends and correlations.From the perspective of the evolution of stock returns dependence,the time-varying Copula model could describe the dynamic dependence between assets appropriately.In this paper,four Internet financial concept shares of different plates are selected as the empirical analysis.Wavelet decomposition is used as a group of low frequency approximation sequences and four sets of high frequency detail sequences.The wavelet coefficients are extracted to fit the optimal class model respectively,and the time-varying Copula model is constructed.Finally,the time-varying correlation graphs of 22 sequence sequences are obtained.When we examine the model,we use the algorithm to reconstruct the correlation coefficient of each decomposition scale and compare it with the original sequence.The study found that(1)there is a significant positive correlation between the mutual gold shares of different plates,among which the strongest correlation is the gold and silver in the middle section.These enterprises reflect the good(or the niche)on the market,indicating that the impact of China's macroeconomic on the plate is consistent;(2)different decomposition scales.There is a certain difference between the Kendall rank correlation coefficient and the original sequence.In general,the low frequency sequence A1 reflects the trend of long-term asset change,the high frequency detail sequence D4 has the largest correlation coefficient and the shortest transaction cycle.(3)the time-varying model can capture the dynamic dependence of each periodic component very well.The tile-varying correlation graph on each decomposition scale is taken as an example.In different degrees of volatility persistence and aggregation,the correlation will change with the time scale of the asset transaction.In general,the dependence of short-term investment is higher than that of the long term.(4)the correlation coefficient map obtained from the reconstruction is exactly the same as the original sequence trend,and the mean square error is only 0.0005,which verifies the wavelet based on the wavelet transform.The scientificity and feasibility of the time-varying Copula model are analyzed.