Symbolic Time Series Analysis Research Articles

Detection of Failure Precursors in Multilayer Ceramic Capacitors Based on Symbolic Time Series Analysis

Detection of Failure Precursors in Multilayer Ceramic Capacitors Based on Symbolic Time Series Analysis

Emergence of Non-predictable Dynamics Caused by Shared Resources in Production Networks

This paper discusses the usage of shared resources in the context of production networks. A simulation study investigates different scenarios of priority concepts for three companies within a small-scale production network including a shared resource. Throughput times of jobs are analyzed by means of statistical distributions as well as symbolic dynamics and time series analysis. It is shown that the collaborative usage of resources induces non-trivial dynamics that are hard to predict. This leads to the conclusion that further research is needed to analyze the possibilities and risks of resource sharing as well as to develop appropriate methods to enable a reasonable usage for manufacturing companies.

Early Detection of Combustion Instability from Hi-speed Flame Images via Deep Learning and Symbolic Time Series Analysis


 
 
 Combustion instability, characterized by self-sustained, large-amplitude pressure oscillations and periodic shedding of coherent vortex structures at varied spatial scales, has many detrimental effects on flight-propulsion dynamics and structural integrity of gas turbine engines. Hence, its early detection is one of the important tasks in engine health monitoring and prognostics. This paper proposes a dynamic data-driven approach, where a large volume of sequential hi-speed (greyscale) images is used to analyze the temporal evolution of coherent structures in combustion chamber for early detection of combustion instability at various operating conditions. The proposed hierarchical approach involves extracting low-dimensional semantic features from images using Deep Neural Networks followed by capturing the temporal evolution of the extracted features using Symbolic Time Series Analysis (STSA). Extensive experimental data have been collected in a swirl-stabilized dump combustor at various operating conditions for validation of the proposed approach. Intermediate layer visualization of deep learning reveals that meaningful shape-features from the flame images are extracted, which enables the temporal modeling layer to enhance the class separability between stable and unstable regions. At the same time, the semantic nature of intermediate features enables expert-guided data exploration that can lead to better understanding of the underlying physics. To the best of the authors knowledge, this paper presents one of the early applications of the recently reported Deep Learning tools in the area of prognostics and health management (PHM).
 
 

Online Detection of Impending Instability in a Combustion System Using Tools from Symbolic Time Series Analysis

In this paper, we introduce a novel technique (anomaly detection) for the online detection of impending instability in a combustion system based on symbolic time series analysis. The experimental results presented in this paper illustrate the application of anomaly detection to a combustor in which the flame is stabilized either by a bluff body or by a swirler. The detection unit works on the principle that in the transition region from combustion noise to thermoacoustic instability, combustion systems exhibit peculiar dynamics which results in the formation of specific patterns in the time series. Further, tools from symbolic time series analysis is used to recognize these patterns and then define an anomaly measure indicative of the proximity of system to regimes of thermoacoustic instability.

Structural breaks and the time-varying levels of weak-form efficiency in crude oil markets: Evidence from the Hurst exponent and Shannon entropy methods

This paper examines the time-varying levels of weak-form efficiency and the presence of structural breaks for two worldwide crude oil benchmarks over the period spanning from January 2, 1990, through September 18, 2012. We use two different econophysics approaches for comparison purposes. The Hurst exponent is provided by the scaled range R/S analysis to measure the degree of long-range dependency exhibited by the West Texas Intermediate (WTI) and European Brent crude oil indices. The Shannon entropy approach, which is based on a symbolic time series analysis (STSA), allows a ranking of market-level efficiency. The empirical results show that the European Brent index is less inefficient than the WTI index for both methods. Moreover, we find that the Hurst exponent displays better performance than the Shannon entropy method. The Hurst exponent is also more effective than the Shannon entropy in detecting financial crashes and crises as well as extreme events, such as wars and terrorist attacks. These findings have several implications for commodity portfolio hedgers and risk managers.

The Fault Pattern Identification of Hydraulic Pump Based on the Sensitive Component STSA

A novel method which is based on ensemble empirical mode decomposition (EEMD) and symbolic time series analysis (STSA) was proposed in this paper. Firstly, the vibration signal of hydraulic pump was decomposed into a number of stationary intrinsic mode functions (IMFs). Secondly, the sensitive component was extracted. Finally, the relative entropy (RE) was extracted from the sensitive components and they were used as the indicator to distinguish the faults of hydraulic pump. The research results of actual testing vibration signal demonstrated the rationality and effectiveness of the proposed method in this paper.

Classification of heart rate signals of healthy and pathological subjects using threshold based symbolic entropy.

The dynamical fluctuations of biological signals provide a unique window to construe the underlying mechanism of the biological systems in health and disease. Recent research evidences suggest that a wide class of diseases appear to degrade the biological complexity and adaptive capacity of the system. Heart rate signals are one of the most important biological signals that have widely been investigated during the last two and half decades. Recent studies suggested that heart rate signals fluctuate in a complex manner. Various entropy based complexity analysis measures have been developed for quantifying the valuable information that may be helpful for clinical monitoring and for early intervention. This study is focused on determining HRV dynamics to distinguish healthy subjects from patients with certain cardiac problems using symbolic time series analysis technique. For that purpose, we have employed recently developed threshold based symbolic entropy to cardiac inter-beat interval time series of healthy, congestive heart failure and atrial fibrillation subjects. Normalized Corrected Shannon Entropy (NCSE) was used to quantify the dynamics of heart rate signals by continuously varying threshold values. A rule based classifier was implemented for classification of different groups by selecting threshold values for the optimal separation. The findings indicated that there is reduction in the complexity of pathological subjects as compared to healthy ones at wide range of threshold values. The results also demonstrated that complexity decreased with disease severity.

Damage localization based on symbolic time series analysis

Summary The objective of this paper is to localize damage in a single or multiple state at early stages of development on the basis of the principles of symbolic dynamics. Symbolic time series analysis (STSA) of noise-contaminated responses is used for feature extraction to detect and localize a gradually evolving deterioration in the structure according to the changes in the statistical behaviour of symbol sequences. Basically, in STSA, statistical features of the symbol sequence can be used to describe the dynamic status of the system. Symbolic dynamics has some useful characteristics making it highly demanded for implementation in real-time observation application such as SHM. First, it significantly reduces the dimension of information and provides information-rich representation of the underlying data. Second, symbolic dynamics and the set of statistical measures built upon it represent a solid framework to address the main challenges of the analysis of nonstationary time data. Finally, STSA often allows capturing the main features of the underlying system whilst alleviating the effects of harmful noise. The method presented in this paper consists of four primary steps: (i) acquisition of the time series data; (ii) creating the symbol space to produce symbol sequences on the basis of the wavelet transformed version of time series data; (iii) developing the symbol probability vectors to achieve anomaly measures; and (iv) localizing damage on the basis of any sudden variation in anomaly measure of different locations. The method was applied on a flexural beam and a 2-D planar truss bridge subjected to varying Gaussian excitation in presence of 2% white noise to examine the efficiency and limitations of the method. Simulation results under various damage conditions confirmed the efficiency of the proposed approach for localization of gradually evolving deterioration in the structure; however, for the future work, the method needs to be verified by experimental data. Copyright © 2014 John Wiley & Sons, Ltd.

Experimental Verification of Parameter Identification Method based on Symbolic Time Series Analysis and Adaptive Immune Clonal Selection Algorithm

The parameter identification method based on symbolic time series analysis (STSA) and adaptive immune clonal selection algorithm (AICSA) was experimentally verified using a 5-story experimental model structure. In the experimental verification, both single and multiple damage scenarios were studied. A 5-story structure was initially healthy with all original columns intact. The single-damage case, the double-damage or the triple-damage case was simulated by replacing the columns of one, two or three different floors, respectively. The experimental results have shown that the parameter identification method based on STSA and AICSA can successfully identify structure parameters only utilizing measured acceleration information for various damage scenarios under different excitation conditions. The proposed approach was shown promising for application of SHM on buildings.

An Independence Test Based on Symbolic Time Series

An independence test based on symbolic time series analysis (STSA) is developed. Considering an independent symbolic time series there is a statistic asymptotically distributed as a CHI-2 with n-1 degrees of freedom. Size and power experiments for small samples were conducted applying Monte Carlo simulations and comparing the results with BDS and runs test. The introduced test shows a good performance detecting independence in nonlinear and chaotic systems.

Experimental and numerical study of the symbolic dynamics of a modulated external-cavity semiconductor laser

We study the symbolic dynamics of a stochastic excitable optical system with periodic forcing. Specifically, we consider a directly modulated semiconductor laser with optical feedback in the low frequency fluctuations (LFF) regime. We use a method of symbolic time-series analysis that allows us to uncover serial correlations in the sequence of intensity dropouts. By transforming the sequence of inter-dropout intervals into a sequence of symbolic patterns and analyzing the statistics of the patterns, we unveil correlations among several consecutive dropouts and we identify clear changes in the dynamics as the modulation amplitude increases. To confirm the robustness of the observations, the experiments were performed using two lasers under different feedback conditions. Simulations of the Lang-Kobayashi (LK) model, including spontaneous emission noise, are found to be in good agreement with the observations, providing an interpretation of the correlations present in the dropout sequence as due to the interplay of the underlying attractor topology, the external forcing, and the noise that sustains the dropout events.

Stock Indices of BRIC economies: Explored for Non Linear Dynamics and Volatility

The current global scenario is a vulnerable one with crisis affecting the global markets. The speed, the magnitude and the frequency of changes has made markets more volatile than ever before. The reports of ECRI (Economic Cycle Research Institute) We are in an era of frequent recession ring true to ears. Predictability in financial markets has long been rejected by the conventional theories of finance. The implicit assumption in these theories that the investors are rational and the financial markets do not possess non linear dynamics has been falsified in many researches. Lately, the tools of Econophysics have proved to successfully overcome this shortcoming. Concepts of Hurst Exponent (Rescaled range analysis) and Shannon Entropy (Symbolic Time Series Analysis) borrowed from Physics and applied in Economics have opened novel dimensions for researchers, policy makers and investors. This paper is an attempt to explore the dynamics of the major stock indices of BRIC economies with these tools. An investigation into existence of the predictability (Hurst Exponent analysis) in these emerging markets coupled with a new volatility measure (Shannon entropy) hint towards non linear dynamics. The findings provide a substantial evidence of foretelling trend reversals in Russian, Indian and Chinese markets. Also, the Indian market was found to be the least volatile amongst the 4 economies. The study would encapsulate wide implications ranging from market efficiency interpretations, to facilitating investors on deciding on their exit and entry into the market and also, to regulators on when to intervene.

Evolving efficiency of the Saudi Arabian stock market via Shannon entropy approach: a study on sector-based data

This paper attempts to analyse the time-varying degrees of weak-for efficiency (WFE) of the Saudi stock market (SSM) using sector-based data. The empirical approach is based on the modified Shannon entropy (SE) and the symbolic time series analysis for a dataset covering 15 sectors for the period (January 2007 to April 2012). Our results reveal significant departure from efficiency but the efficiency degree evolves over time. Also, we show the existence of structural breaks in the time-varying measure of efficiency indicating that the occurrence of extreme events strongly impacts the efficiency behaviour. Furthermore, we provide evidence that Shariah-compliant stocks behave differently from other firms in terms of the degree of WFE and its path over time. Overall, our findings provide several implications for portfolio managers and Saudi policy makers.

Characterizing the dynamics of coupled pendulums via symbolic time series analysis

We propose a novel method of symbolic time-series analysis aimed at characterizing the regular or chaotic dynamics of coupled oscillators. The method is applied to two identical pendulums mounted on a frictionless platform, resembling Huygens’ clocks. Employing a transformation rule inspired in ordinal analysis [C. Bandt and B. Pompe, Phys. Rev. Lett. 88, 174102 (2002)], the dynamics of the coupled system is represented by a sequence of symbols that are determined by the order in which the trajectory of each pendulum intersects an appropriately chosen hyperplane in the phase space. For two coupled pendulums we use four symbols corresponding to the crossings of the vertical axis (at the bottom equilibrium point), either clock-wise or anti-clock wise. The complexity of the motion, quantified in terms of the entropy of the symbolic sequence, is compared with the degree of chaos, quantified in terms of the largest Lyapunov exponent. We demonstrate that the symbolic entropy sheds light into the large variety of different periodic and chaotic motions, with different types synchronization, that cannot be inferred from the Lyapunov analysis.

Abnormal state detection of building structures based on symbolic time series analysis and negative selection

A symbolization-based negative selection (SNS) algorithm that combines the advantages of symbolic time series analysis (STSA) and negative selection (NS) is proposed for detecting the abnormal states of building structures. In SNS, no prior knowledge of a structure's abnormal state is needed. Only the response of the structure in a current state is used as input data. In addition, this approach can work even with one sensor, so it is highly practical and flexible. A state sequence histogram (SSH) transformed from raw acceleration data by using STSA can capture the main features of structure dynamics and alleviate the effects of harmful noise. SSHs of the normal and abnormal states of a structure are defined as self and non-self elements, respectively. A new detector generation strategy and matching mechanism is proposed that makes the procedure more effective, along with guidelines for appropriate parameters in SNS. Numerical simulations and experimental verifications for different abnormal state cases were conducted to demonstrate the feasibility of the proposed method. Copyright © 2013 John Wiley & Sons, Ltd.

Extracting ordinal temporal trail clusters in networks using symbolic time-series analysis

Temporal trails generated by agents traveling to various locations at different time epochs are becoming more prevalent in large social networks. We propose an algorithm to intuitively cluster groups of such agent trails from networks. The proposed algorithm is based on modeling each trail as a probabilistic finite state automata (PFSA). The algorithm also allows the specification of the required degree of similarity between the trails by specifying the depth of the PFSA. Hierarchical agglomerative clustering is used to group trails based on their representative PFSA and the locations that they visit. The algorithm was applied to simulated trails and real-world network trails obtained from merchant marine ships GPS locations. In both cases it was able to intuitively detect and extract the underlying patterns in the trails and form clusters of similar trails.

Hybrid Methodology for Structural Health Monitoring Based on Immune Algorithms and Symbolic Time Series Analysis

This hybrid methodology for structural health monitoring (SHM) is based on immune algorithms (IAs) and symbolic time series analysis (STSA). Real-valued negative selection (RNS) is used to detect damage detection and adaptive immune clonal selection algorithm (AICSA) is used to localize and quantify the damage. Data symbolization by using STSA alleviates the effects of harmful noise in raw acceleration data. This paper explains the mathematical basis of STSA and the procedure of the hybrid methodology. It also describes the results of an simulation experiment on a five-story shear frame structure that indicated the hybrid strategy can efficiently and precisely detect, localize and quantify damage to civil engineering structures in the presence of measurement noise.

Symbolization-based differential evolution strategy for identification of structural parameters

SUMMARY This new method of identifying structural parameters, called ‘symbolization-based differential evolution strategy’ (SDES), merges the advantages of symbolic time series analysis and differential evolution (DE). Data symbolization in SDES alleviates the effects of harmful noise. SDES was numerically compared with particle swarm optimization and DE on raw acceleration data. These simulations revealed that SDES provided better estimates of structural parameters when the data were contaminated by noise. We applied SDES to experimental data to assess its feasibility in realistic problems. SDES performed much better than particle swarm optimization and DE on raw acceleration data. The simulations and experiments show that SDES is a powerful tool for identifying unknown parameters of structural systems even when the data are contaminated with relatively large amounts of noise. Copyright © 2012 John Wiley & Sons, Ltd.

Crash occurrence probability and stock market efficiency the indie stock exchange case via Shannon entropy

In this study, we evaluate the relationship between efficiency and probability of the crash, thus the evolution of the daily informational efficiency is measured for the indie stock market index. The efficiency, which is the issue addressed by the weak-form efficient market hypothesis, is calculated using a new method the Shannon entropy and the symbolic time series analysis. A logit model is applied in order to study the relationship between efficiency and probability of the financial crash.

Feasibility Study of Parameter Identification Method Based on Symbolic Time Series Analysis and Adaptive Immune Clonal Selection Algorithm

The feasibility of a parameter identification method based on symbolic time series analysis (STSA) and the adaptive immune clonal selection algorithm (AICSA) is studied. Data symbolization by using STSA alleviates the effects of harmful noise in raw acceleration data. The effect of the parameters in STSA is theoretically evaluated and numerically verified. AICSA is employed to minimize the error between the state sequence histogram (SSH) that is transformed from raw acceleration data by STSA. The proposed methodology is evaluated by comparing it with AICSA using raw acceleration data. AICSA combining STSA is proved to be a powerful tool for identifying unknown parameters of structural systems even when the data is contaminated with relatively large amounts of noise.