Multiscale Detrended Cross-correlation Analysis Research Articles

Damage detection of structures with detrended fluctuation and detrended cross-correlation analyses

Recently, fractal analysis has shown its potential for damage detection and assessment in fields such as biomedical and mechanical engineering. For its practicability in interpreting irregular, complex, and disordered phenomena, a structural health monitoring (SHM) system based on detrended fluctuation analysis (DFA) and detrended cross-correlation analysis (DCCA) is proposed. First, damage conditions can be swiftly detected by evaluating ambient vibration signals measured from a structure through DFA. Damage locations can then be determined by analyzing the cross correlation of signals of different floors by applying DCCA. A damage index is also proposed based on multi-scale DCCA curves to improve the damage location accuracy. To verify the performance of the proposed SHM system, a four-story numerical model was used to simulate various damage conditions with different noise levels. Furthermore, an experimental verification was conducted on a seven-story benchmark structure to assess the potential damage. The results revealed that the DFA method could detect the damage conditions satisfactorily, and damage locations can be identified through the DCCA method with an accuracy of 75%. Moreover, damage locations can be correctly assessed by the damage index method with an improved accuracy of 87.5%. The proposed SHM system has promising application in practical implementations.

Multiscale Detrended Cross-Correlation Analysis of Traffic Time Series Based on Empirical Mode Decomposition

In this paper, we propose multiscale detrended cross-correlation analysis (MSDCCA) to detect the long-range power-law cross-correlation of considered signals in the presence of nonstationarity. For improving the performance and getting better robustness, we further introduce the empirical mode decomposition (EMD) to eliminate the noise effects and propose MSDCCA method combined with EMD, which is called MS-EDXA method, then systematically investigate the multiscale cross-correlation structure of the real traffic signals. We apply the MSDCCA and MS-EDXA methods to study the cross-correlations in three situations: velocity and volume on one lane, velocities on the present and the next moment and velocities on the adjacent lanes, and further compare their spectrums respectively. When the difference between the spectrums of MSDCCA and MS-EDXA becomes unobvious, there is a crossover which denotes the turning point of difference. The crossover results from the competition between the noise effects in the original signals and the intrinsic fluctuation of traffic signals and divides the plot of spectrums into two regions. In all the three case, MS-EDXA method makes the average of local scaling exponents increased and the standard deviation decreased and provides a relative stable persistent scaling cross-correlated behavior which gets the analysis more precise and more robust and improves the performance after noises being removed. Applying MS-EDXA method avoids the inaccurate characteristics of multiscale cross-correlation structure at the short scale including the spectrum minimum, the range for the spectrum fluctuation and general trend, which are caused by the noise in the original signals. We get the conclusions that the traffic velocity and volume are long-range cross-correlated, which is accordant to their actual evolution, while velocities on the present and the next moment and velocities on adjacent lanes reflect the strong cross-correlations both in temporal and spatial dimensions. We also reveal the similarity and uniqueness in the cross-correlation situations between velocities. Besides, signals on one lane show stronger long-range cross-correlation than that on adjacent lanes. Thus, the multiscale cross-correlation structure acquired by MS-EDXA is more close to the intrinsic mechanism of traffic system and reflects more accurate and more abundant traffic information.

MULTISCALE DETRENDED CROSS-CORRELATION ANALYSIS OF STOCK MARKETS

In this paper, we employ the detrended cross-correlation analysis (DCCA) to investigate the cross-correlations between different stock markets. We report the results of cross-correlated behaviors in US, Chinese and European stock markets in period 1997–2012 by using DCCA method. The DCCA shows the cross-correlated behaviors of intra-regional and inter-regional stock markets in the short and long term which display the similarities and differences of cross-correlated behaviors simply and roughly and the persistence of cross-correlated behaviors of fluctuations. Then, because of the limitation and inapplicability of DCCA method, we propose multiscale detrended cross-correlation analysis (MSDCCA) method to avoid "a priori" selecting the ranges of scales over which two coefficients of the classical DCCA method are identified, and employ MSDCCA to reanalyze these cross-correlations to exhibit some important details such as the existence and position of minimum, maximum and bimodal distribution which are lost if the scale structure is described by two coefficients only and essential differences and similarities in the scale structures of cross-correlation of intra-regional and inter-regional markets. More statistical characteristics of cross-correlation obtained by MSDCCA method help us to understand how two different stock markets influence each other and to analyze the influence from thus two inter-regional markets on the cross-correlation in detail, thus we get a richer and more detailed knowledge of the complex evolutions of dynamics of the cross-correlations between stock markets. The application of MSDCCA is important to promote our understanding of the internal mechanisms and structures of financial markets and helps to forecast the stock indices based on our current results demonstrated the cross-correlations between stock indices. We also discuss the MSDCCA methods of secant rolling window with different sizes and, lastly, provide some relevant implications and issue.

Asymmetric multiscale detrended cross-correlation analysis of financial time series

We propose the asymmetric multiscale detrended cross-correlation analysis (MS-ADCCA) method and apply MS-ADCCA method to explore the existence of asymmetric cross-correlation for daily price returns in US and Chinese stock markets and to assess the properties of these asymmetric cross-correlations. The results all show the existences of asymmetric cross-correlations, while small asymmetries at small scales and larger asymmetries at larger scales are also displayed. There is a strong similarity between S&P500 and DJI, and we reveal that the asymmetries depend more on the cross-correlations of S&P500 vs. DJI, S&P500 vs. NQCI, DJI vs. NQCI, and ShangZheng vs. ShenCheng when the market is falling than rising, respectively. By comparing the spectra of S&P500 vs. NQCI and DJI vs. NQCI with uptrends and downtrends, we detect some new characteristics which lead to some new conclusions. Likewise, some new conclusions also can be drawn by the new characteristics displayed through the comparison between the spectra of ShangZheng vs. HSI and ShenCheng vs. HSI. Obviously, we conclude that although the overall spectra are similar and one market has the same effect when it is rising and falling in the study of asymmetric cross-correlations between it and different markets, the cross-correlations and asymmetries on the trends of the different markets are all different. MS-ADCCA method can detect the differences on the asymmetric cross-correlations by different trends of markets. Moreover, the uniqueness of cross-correlation between NQCI and HSI can be detected in the study of the asymmetric cross-correlations, which confirms that HSI is unique in the Chinese stock markets and NQCI is unique in the US stock markets further.

Comparison of Multiscale Methods in the Stock Markets for Detrended Cross-correlation Analysis and Cross-sample Entropy

In this paper, we employ multiscale cross-sample entropy (MSCE), multiscale detrended cross-correlation analysis (MSDCCA) and DCCA cross-correlation coefficient (σDCCA) measurement to investigate the relationship between time series among different stock markets. We report the results of synchronism and cross-correlation behaviors in US and Chinese stock markets by these three methods. It can be concluded that the MSCE analysis point out the similarity on the cross-correlation among the stock markets while the MSCE makes it difficult to distinguish the indices in the same region and identify the difference and uniqueness of stock markets. However, both the MSDCCA analysis and σDCCA analysis reflect the similarity and uniqueness on the cross-correlation behaviors and reach the consistency. Furthermore, MSDCCA gives detailed multiscale cross-correlation structures and show some new interesting characteristics and conclusions, while the multiscale analysis by σDCCA provides a large amount of information on the cross-correlations and quantifies the level of cross-correlation more clearly and intuitively. MSDCCA and σDCCA methods may be more proper measures for the investigation of the cross-correlation between time series. We believe that such researches are relevant for a better understanding of the stock market mechanisms, and may lead to a better forecasting of the stock indices.

Modified DFA and DCCA approach for quantifying the multiscale correlation structure of financial markets

We use multiscale detrended fluctuation analysis (MSDFA) and multiscale detrended cross-correlation analysis (MSDCCA) to investigate auto-correlation (AC) and cross-correlation (CC) in the US and Chinese stock markets during 1997–2012. The results show that US and Chinese stock indices differ in terms of their multiscale AC structures. Stock indices in the same region also differ with regard to their multiscale AC structures. We analyze AC and CC behaviors among indices for the same region to determine similarity among six stock indices and divide them into four groups accordingly. We choose S&P500, NQCI, HSI, and the Shanghai Composite Index as representative samples for simplicity. MSDFA and MSDCCA results and average MSDFA spectra for local scaling exponents (LSEs) for individual series are presented. We find that the MSDCCA spectrum for LSE CC between two time series generally tends to be greater than the average MSDFA LSE spectrum for individual series. We obtain detailed multiscale structures and relations for CC between the four representatives. MSDFA and MSDCCA with secant rolling windows of different sizes are then applied to reanalyze the AC and CC. Vertical and horizontal comparisons of different window sizes are made. The MSDFA and MSDCCA results for the original window size are confirmed and some new interesting characteristics and conclusions regarding multiscale correlation structures are obtained.