Diagnosis and impacts of non-Gaussianity of innovations in data assimilation

Most of the atmospheric and oceanic data assimilation (DA) schemes rely on the Best Linear Unbiased Estimator (BLUE), which is sub-optimal if errors of assimilated data are non-Gaussian, thus calling for a full Bayesian data assimilation. This paper contributes to the study of the non-Gaussianity of errors in the observational space. Possible sources of non-Gaussianity range from the inherent statistical skewness and positiveness of some physical observables (e.g. moisture, chemical species), the nonlinearity, both of the data assimilation models and of the observation operators among others. Deviations from Gaussianity can be justified from a priori hypotheses or inferred from statistical diagnostics of innovations (observation minus background), leading to consistency relationships between the error statistics. From samples of observations and backgrounds as well as their specified error variances, we evaluate some measures of the innovation non-Gaussianity, such as the skewness, kurtosis and negentropy. Under the assumption of additive errors and by relating statistical moments from both data errors and innovations, we identify potential sources of the innovation non-Gaussianity. These sources range from: (1) univariate error non-Gaussianity, (2), nonlinear correlations between errors, (3) spatio-temporal variability of error variances (heteroscedasticity) and (4) multiplicative noise. Observational and background errors are often assumed independent. This leads to variance-dependent bounds for the skewness and the kurtosis of errors. From innovation statistics, we assess the potential DA impact of some scenarios of non-Gaussian errors. This impact is measured through the mean square difference between the BLUE and the Minimum Variance Unbiased Estimator (MVUE), obtained with univariate observations and background estimates. In order to accomplish this, we compute maximum entropy probability density functions (pdfs) of the errors, constrained by the first four order moments. These pdfs are then used to compute the Bayesian posterior pdf and the MVUE. The referred impact is studied for a large range of statistical moments, being higher for skewed innovations and growing in average with the skewness of data errors, specially if the skewnesses have the same sign. An application has been performed to the quality-accepted ECMWF innovations of brightness temperatures of a set of High Resolution Infrared Sounder (HIRS) channels. In this context, the MVUE has led in some extreme cases to a potential reduction of 20%-60% of the posterior error variance as compared to the BLUE, specially for extreme values of the innovations.     

Bayesian and Frequentist Inferences on a Type I Half-Logistic Odd Weibull Generator with Applications in Engineering

In this article, we have proposed a new generalization of the odd Weibull-G family by consolidating two notable families of distributions. We have derived various mathematical properties of the proposed family, including quantile function, skewness, kurtosis, moments, incomplete moments, mean deviation, Bonferroni and Lorenz curves, probability weighted moments, moments of (reversed) residual lifetime, entropy and order statistics. After producing the general class, two of the corresponding parametric statistical models are outlined. The hazard rate function of the sub-models can take a variety of shapes such as increasing, decreasing, unimodal, and Bathtub shaped, for different values of the parameters. Furthermore, the sub-models of the introduced family are also capable of modelling symmetric and skewed data. The parameter estimation of the special models are discussed by numerous methods, namely, the maximum likelihood, simple least squares, weighted least squares, Cramér-von Mises, and Bayesian estimation. Under the Bayesian framework, we have used informative and non-informative priors to obtain Bayes estimates of unknown parameters with the squared error and generalized entropy loss functions. An extensive Monte Carlo simulation is conducted to assess the effectiveness of these estimation techniques. The applicability of two sub-models of the proposed family is illustrated by means of two real data sets.     

A New Generator of Probability Models: The Exponentiated Sine-G Family for Lifetime Studies

In this article, we propose the exponentiated sine-generated family of distributions. Some important properties are demonstrated, such as the series representation of the probability density function, quantile function, moments, stress-strength reliability, and Rényi entropy. A particular member, called the exponentiated sine Weibull distribution, is highlighted; we analyze its skewness and kurtosis, moments, quantile function, residual mean and reversed mean residual life functions, order statistics, and extreme value distributions. Maximum likelihood estimation and Bayes estimation under the square error loss function are considered. Simulation studies are used to assess the techniques, and their performance gives satisfactory results as discussed by the mean square error, confidence intervals, and coverage probabilities of the estimates. The stress-strength reliability parameter of the exponentiated sine Weibull model is derived and estimated by the maximum likelihood estimation method. Also, nonparametric bootstrap techniques are used to approximate the confidence interval of the reliability parameter. A simulation is conducted to examine the mean square error, standard deviations, confidence intervals, and coverage probabilities of the reliability parameter. Finally, three real applications of the exponentiated sine Weibull model are provided. One of them considers stress-strength data.     

Existence conditions and spreading properties of extreme entropy D-dimensional distributions

The extremization of the information-theoretic measures (Fisher information, Shannon entropy, Tsallis entropy), which complementary describe the spreading of the physical states of natural systems, gives rise to fundamental equations of motion and/or conservation laws. At times, the associated extreme entropy distributions are known for some given constraints, usually moments or radial expectation values. In this work, first we give the existence conditions of the maxent probability distributions in a D-dimensional scenario where two moments (not necessarily of consecutive order) are known. Then we find general relations which involve four elements (the extremized entropy, the other two information-theoretic measures and the variance of the extremum density) in scenarios with different dimensionalities and moment constraints.     

The Role of a Thick Soliton in the Dynamics of the Soliton Gas Within the Framework of the Gardner Equation

Radiophysics and Quantum Electronics - We study the statistical moments of the soliton gas (mean field, variance, skewness, and kurtosis), which is described within the framework of the Gardner...     

An econophysics approach to analyse uncertainty in financial markets: an application to the Portuguese stock market

In recent years there has been a closer interrelationship between several scientific areas trying to obtain a more realistic and rich explanation of the natural and social phenomena. Among these it should be emphasized the increasing interrelationship between physics and financial theory. In this field the analysis of uncertainty, which is crucial in financial analysis, can be made using measures of physics statistics and information theory, namely the Shannon entropy. One advantage of this approach is that the entropy is a more general measure than the variance, since it accounts for higher order moments of a probability distribution function. An empirical application was made using data collected from the Portuguese Stock Market.     

Two-soliton interaction as an elementary act of soliton turbulence in integrable systems

Two-soliton interactions play a definitive role in the formation of the structure of soliton turbulence in integrable systems. To quantify the contribution of these interactions to the dynamical and statistical characteristics of the nonlinear wave field of soliton turbulence we study properties of the spatial moments of the two-soliton solution of the Korteweg–de Vries (KdV) equation. While the first two moments are integrals of the KdV evolution, the 3rd and 4th moments undergo significant variations in the dominant interaction region, which could have strong effect on the values of the skewness and kurtosis in soliton turbulence.     

基于超声射频信号的支持向量机双参量B线识别方法*

肺超声中的特殊征象B线对于临床诊断肺水肿等肺部疾病有重要意义,但诊断结果依赖于医生的主观判断,为了客观、自动地识别B线,提高诊断准确率,本文提出了一种基于超声回波射频信号的肺脏超声特殊征象B线识别方法。本文首先选取了射频信号的排列熵、信息熵、峰度、偏度、能量作为特征参数,利用独立样本t检验和单参数贝叶斯分类的方法检验超声射频数据中B线以及非B线所对应射频数据的各个参量的差异性以及各参数与B线识别的相关性。然后将不同的双参量组合输入非线性支持向量机（SVM）中进行分类,比较各个组合的分类效果。结果显示信息熵与排列熵参数组合基于射频信号的分类效果最好,分类灵敏度为90.521%,特异性为98.106%,准确率为96.328%,AUC等于0.95。在引入后处理算法后,B线识别效果有进一步提升,得到分类平均灵敏度为95.23%,平均特异性为97.22%,平均准确率为96.88%。研究结果表明基于射频数据的SVM双参量B线识别方法对辅助临床诊断具有重要价值,信息熵和排列熵的组合可以有效的对特殊征象B线进行高精度识别。     

Piston scuffing fault and its identification in an IC engine by vibration analysis

In this paper, the effects of piston scuffing fault on engine performance and vibrations are investigated. A procedure based on vibration analysis is also presented to identify piston scuffing fault. To this end, an internal combustion (IC) engine ran under a specific test procedure. The engine parameters and vibration signals were measured during the experiments. To produce piston scuffing fault, three-body abrasive wear mechanism was employed. The experimental results showed that piston scuffing fault caused the engine performance to reduce significantly. The vibration signals were analyzed in time-domain, frequency-domain and time–frequency domain. Continuous wavelet transform (CWT) was used to obtain time–frequency representations. “dmey” wavelet was selected as the optimum wavelet type for this research among different wavelet types using the three criteria of energy, Shannon entropy and energy to Shannon entropy ratio. The results of CWT analysis by “dmey” wavelet showed that piston scuffing fault excited the frequency band of 2.4–4.7 kHz in which the frequency of 3.7 kHz was affected more. Finally, seven different features were extracted from the engine vibration signals related to the frequency band of 2.4–4.7 kHz. The results indicated that maximum, mean, RMS, skewness, kurtosis and impulse factor of the engine vibration related to the found frequency band increased significantly due to piston scuffing fault. The obtained results showed that the proposed method identified piston scuffing fault and discovered the vibration characteristics of this fault like frequency band. The results also demonstrated the possibility of using engine vibrations in piston scuffing fault identification.     

Shannon Entropy in LS-Coupled Configuration Space for Ni-like Isoelectronic Sequence

The Shannon entropy in an LS-coupled configuration space has been calculated through a transformation from that in a jj-coupled configuration space for a Ni-like isoelectronic sequence. The sudden change of Shannon entropy, information exchange, eigenlevel anticrossing, and strong configuration interaction have been presented for adjacent levels. It is shown that eigenlevel anticrossing is a sufficient and necessary condition for the sudden change of Shannon entropy, and both of them are a sufficient condition for information exchange, which is the same as the case of the jj-coupled configuration space. It is found that the structure of sudden change from jj-coupled into LS-coupled configuration spaces through the LS-jj transformation is invariant for Shannon entropy along the isoelectronic sequence. What is more, in an LS-coupled configuration space, there are a large number of information exchanges between energy levels whether with or without strong configuration interaction, and most of the ground and single excited states of Ni-like ions are more suitable to be described by a jj-coupled or other configuration basis set instead of an LS-coupled configuration basis set according to the configuration mixing coefficients and their Shannon entropy. In this sense, Shannon entropy can also be used to measure the applicability of a configuration basis set or the purity of atomic state functions in different coupling schemes.     

Spectral Moments of the Long-term Average Spectrum: Sensitive Indices of Voice Change After Therapy?

Voice clinicians require an objective, reliable, and relatively automatic method to assess voice change after medical, surgical, or behavioral intervention. This measure must be sensitive to a variety of voice qualities and severities, and preferably should reflect voice in continuous speech. The long-term average spectrum (LTAS) is a fast Fourier transform-generated power spectrum whose properties can be compared with a Gaussian bell curve using spectral moments analysis. Four spectral moments describe features of the LTAS: Spectral mean (Moment 1) and standard deviation (Moment 2) represent the spectrum's central tendency and dispersion, respectively. Skewness (based on Moment 3) and kurtosis (based on Moment 4) represent the spectrum's tilt and peakedness, respectively. To examine whether the first four spectral moments of the LTAS were sensitive to perceived voice improvement after voice therapy, this investigation compared pretreatment and posttreatment voice samples of 93 patients with functional dysphonia using spectral moments analysis. Inspection of the results revealed that spectral mean and standard deviation lowered significantly with perceived voice improvement after successful behavioral management (p < 0.001). However, changes in skewness and kurtosis were not significant. Furthermore, lowering of the spectral mean uniquely accounted for approximately 14% of the variance in the pretreatment to posttreatment changes observed in perceptual ratings of voice severity (p < 0.001), indicating that spectral mean (ie, Moment 1) of the LTAS may be one acoustic marker sensitive to improvement in dysphonia severity.     

基于概率密度分布型变化的突变检测新途径

对于一个稳定的动力系统而言,系统变量的概率密度具有较为稳定的分布型,而当系统的动力学结构发生变化后可能会导致系统变量的分布型发生不同程度的变化.鉴于此,本文从识别系统变量的概率密度分布的微小变化角度出发,将描述时间序列概率分布特征的偏度系数和峰度系数应用于时间序列的突变检测中.数值试验结果表明,偏度系数和峰度系数对突变信号具有很好的识别能力,进而揭示了一条检测突变的新途径.进一步的研究表明,新方法的检测结果对于子序列长度的选择具有较小的依赖性.     

Recurrence plots and Shannon entropy for a dynamical analysis of asynchronisms in noninvasive mechanical ventilation

Recurrence plots were introduced to quantify the recurrence properties of chaotic dynamics. Hereafter, the recurrence quantification analysis was introduced to transform graphical interpretations into statistical analysis. In this spirit, a new definition for the Shannon entropy was recently introduced in order to have a measure correlated with the largest Lyapunov exponent. Recurrence plots and this Shannon entropy are thus used for the analysis of the dynamics underlying patient assisted with a mechanical noninvasive ventilation. The quality of the assistance strongly depends on the quality of the interactions between the patient and his ventilator which are crucial for tolerance and acceptability. Recurrence plots provide a global view of these interactions and the Shannon entropy is shown to be a measure of the rate of asynchronisms as well as the breathing rhythm.     

Probing the QCD critical point by higher moments of net-proton multiplicity distributions at STAR

Higher moments (kurtosis (??), skewness (S) and variance (?? ²)) of multiplicity distributions are sensitive to the correlation length and can be used to search for the QCD critical point. The moment products ??? ² and S ?? of net-proton distributions, which are also related to volume independent baryon number susceptibility ratios, can be compared with Lattice QCD and Hadron Resonance Gas (HRG) model calculations. We discuss the recent progress in the higher moments analysis of net-protons multiplicity distributions.     

Regularities unseen,randomness observed: levels of entropy convergence

We study how the Shannon entropy of sequences produced by an information source converges to the source's entropy rate. We synthesize several phenomenological approaches to applying information theoretic measures of randomness and memory to stochastic and deterministic processes by using successive derivatives of the Shannon entropy growth curve. This leads, in turn, to natural measures of apparent memory stored in a source and the amounts of information that must be extracted from observations of a source in order for it to be optimally predicted and for an observer to synchronize to it. To measure the difficulty of synchronization, we define the transient information and prove that, for Markov processes, it is related to the total uncertainty experienced while synchronizing to a process. One consequence of ignoring a process's structural properties is that the missed regularities are converted to apparent randomness. We demonstrate that this problem arises particularly for settings where one has access only to short measurement sequences. Numerically and analytically, we determine the Shannon entropy growth curve, and related quantities, for a range of stochastic and deterministic processes. We conclude by looking at the relationships between a process's entropy convergence behavior and its underlying computational structure.     

非定常输运问题适应于消息传递并行编程环境的香农熵计算方法

在多计算步的非定常输运问题的蒙特卡罗模拟中,为自动调整每一步的样本数以获得较高的计算效率,可以有多种准则.一种可选的方法是在每一步每隔若干样本监测一次系统中未死亡粒子属性分布对应的香农熵的收敛情况以决定何时停止追加样本,此种方法需要在每一步频繁计算香农熵值.由于在MPI消息传递并行编程环境下香农熵的经典计算方法必须广播大量的数据,导致每一步的计算时间随香农熵计算频率的提高而快速增大,这显然是不能满足实际需求的.本文提出了一种适应于消息传递并行编程环境的香农熵计算新方法,该方法计算得到的香农熵值并不等价于经典方法,但二者之间的差别会随着样本数的增加而趋于零.新方法的最大优势是高频计算香农熵值的时间代价大为降低,为最终实现基于香农熵收敛判断的每步样本数的自动调整奠定了必要的基础.     

Brownian Particle in a Poisson-Shot-Noise Active Bath: Exact Statistics,Effective Temperature,and Inference

The dynamics of an overdamped Brownian particle in a thermal bath that contains a dilute solution of active particles is studied. The particle moves in a harmonic potential and experiences Poisson shot-noise kicks with specified amplitude distribution due to moving active particles in the bath. From the Fokker–Planck equation for the particle dynamics, the stationary solution for the displacement distribution is derived along with the moments characterizing mean, variance, skewness, and kurtosis, as well as finite-time first and second moments. An effective temperature is also computed through the fluctuation–dissipation theorem and show that equipartition theorem holds for all zero-mean kick distributions, including those leading to non-Gaussian stationary statistics. For the case of Gaussian-distributed active kicks, a re-entrant behavior from non-Gaussian to Gaussian stationary states and a heavy-tailed leptokurtic distribution across a wide range of parameters are found as seen in recent experimental studies. Further analysis reveals statistical signatures of the irreversible dynamics of the particle displacement in terms of the time asymmetry of cross-correlation functions. Fruits of the work is the development of an compact inference scheme that may allow experimentalists to extract the rate and moments of underlying shot-noise solely from the statistics the particle position.     

Application of entropies to the study of the decoherence of magnetopolaron in 0-D nanosystem

Many methods have been experimented to study decoherence in quantum dot (QD). Tsallis, Shannon and Gaussian entropy have been used to study decoherence separately; in this paper, we compared the results of the Gaussian, Shannon, and Tsallis entropies in 0-D nanosystem. The linear combination operator and the unitary transformation was used to derive the magnetopolaron spectrum that strongly interacts with the LO phonons in the presence of an electric field in the pseudoharmonic and delta quantum dot. Numerical results revealed for the quantum pseudo dot that: (i) the amplitude of Gauss entropy is greater than the amplitude of Tsallis entropy which in turn is greater than the amplitude of Shannon entropy. The Tsallis entropy is not more significant in nanosystem compared to Shannon and Gauss entropies, (ii) with an increase of the zero point, the dominance of the Gauss entropy on the Shannon entropy was observed on one hand and the dominance of the Shannon entropy on the Tsallis entropy on the other hand; this suggested that in nanosystem, Gauss entropy is more suitable in the evaluation of the average of information in the system, for the delta quantum dot it was observed that (iii) when the Gauss entropy is considered, a lot of information about the system is missed. The collapse revival phenomenon in Shannon entropy was observed in RbCl and GaAs delta quantum dot with the enhancement of delta parameter; with an increase in this parameter, the system in the case of CsI evolved coherently; with Shannon and Tsallis entropies, information in the system is faster and coherently exchanged; (iv) the Shannon entropy is more significant because its amplitude outweighs the others when the delta dimension length enhances. The Tsallis entropy involves as wave bundle; which oscillate periodically with an increase of the oscillation period when delta dimension length is improved.     

Numerical Algorithms for Estimating Probability Density Function Based on the Maximum Entropy Principle and Fup Basis Functions

Estimation of the probability density function from the statistical power moments presents a challenging nonlinear numerical problem posed by unbalanced nonlinearities, numerical instability and a lack of convergence, especially for larger numbers of moments. Despite many numerical improvements over the past two decades, the classical moment problem of maximum entropy (MaxEnt) is still a very demanding numerical and statistical task. Among others, it was presented how Fup basis functions with compact support can significantly improve the convergence properties of the mentioned nonlinear algorithm, but still, there is a lot of obstacles to an efficient pdf solution in different applied examples. Therefore, besides the mentioned classical nonlinear Algorithm 1, in this paper, we present a linear approximation of the MaxEnt moment problem as Algorithm 2 using exponential Fup basis functions. Algorithm 2 solves the linear problem, satisfying only the proposed moments, using an optimal exponential tension parameter that maximizes Shannon entropy. Algorithm 2 is very efficient for larger numbers of moments and especially for skewed pdfs. Since both Algorithms have pros and cons, a hybrid strategy is proposed to combine their best approximation properties.