Hybrid linear and nonlinear complexity pursuit for blind source separation

Blind source separation (BSS) is an increasingly popular data analysis technique with many applications. Several methods for BSS using the statistical properties of original sources have been proposed; for a famous case, non-Gaussianity, this leads to independent component analysis (ICA). In this paper, we propose a hybrid BSS method based on linear and nonlinear complexity pursuit, which combines three statistical properties of source signals: non-Gaussianity, linear predictability and nonlinear predictability. A gradient learning algorithm is presented by minimizing a loss function. Simulations verify the efficient implementation of the proposed method.     

Blind source separation with nonlinear autocorrelation and non-Gaussianity

Blind source separation (BSS) is a problem that is often encountered in many applications, such as biomedical signal processing and analysis, speech and image processing, wireless telecommunication systems, data mining, sonar, radar enhancement, etc. One often solves the BSS problem by using the statistical properties of original sources, e.g., non-Gaussianity or time-structure information. Nevertheless, real-life mixtures are likely to contain both non-Gaussianity and time-structure information sources, rendering the algorithms using only one statistical property fail. In this paper, we address the BSS problem when source signals have non-Gaussianity and temporal structure with nonlinear autocorrelation. Based on the two statistical characteristics of sources, we develop an objective function. Maximizing the objective function, we propose a gradient ascent source separation algorithm. Furthermore, We give some mathematical properties for the algorithm. Computer simulations for sources with square temporal autocorrelation and non-Gaussianity illustrate the efficiency of the proposed approach.     

A fixed-point algorithm for blind source separation with nonlinear autocorrelation

This paper addresses blind source separation (BSS) problem when source signals have the temporal structure with nonlinear autocorrelation. Using the temporal characteristics of sources, we develop an objective function based on the nonlinear autocorrelation of sources. Maximizing the objective function, we propose a fixed-point source separation algorithm. Furthermore, we give some mathematical properties of the algorithm. Computer simulations for sources with square temporal autocorrelation and the real-world applications in the analysis of the magnetoencephalographic recordings (MEG) illustrate the efficiency of the proposed approach. Thus, the presented BSS algorithm, which is based on the nonlinear measure of temporal autocorrelation, provides a novel statistical property to perform BSS.     

基于规范高阶累积量的盲分离算法

本文给出信号的r阶规范累积量定义,证明了在信号瞬时线性混叠情况下r阶规范累积量绝对值不超过最大源信号的r阶规范累积量,因此可以通过最大化r阶规范累积量的绝对值实现盲分离.最大化r阶规范累积量的绝对值可以得到两种盲分离算法,一种高效快速算法———特征值分解盲分离算法,另外一种盲提取算法.本文为高阶累积量盲分离奠定了严格的理论基础.仿真结果验证了理论的正确性和算法的有效性.     

Blind separation of vibration signals and source change detection – Application to machine monitoring

The paper presents a new approach for machine vibration analysis and health monitoring combining blind source separation (BSS) and change detection in source signals. So, the problem is transferred from the original space of the measurements to the space of independent sources, where the reduced number of components is going to simplify the monitoring problem while the change detection methods are going to be applied for scalar signals. The assessment of the approach on a real machine is presented in this paper.     

A non-negative matrix factorization model based on the zero-inflated Tweedie distribution

Non-negative matrix factorization (NMF) is a technique of multivariate analysis used to approximate a given matrix containing non-negative data using two non-negative factor matrices that has been applied to a number of fields. However, when a matrix containing non-negative data has many zeroes, NMF encounters an approximation difficulty. This zero-inflated situation occurs often when a data matrix is given as count data, and becomes more challenging with matrices of increasing size. To solve this problem, we propose a new NMF model for zero-inflated non-negative matrices. Our model is based on the zero-inflated Tweedie distribution. The Tweedie distribution is a generalization of the normal, the Poisson, and the gamma distributions, and differs from each of the other distributions in the degree of robustness of its estimated parameters. In this paper, we show through numerical examples that the proposed model is superior to the basic NMF model in terms of approximation of zero-inflated data. Furthermore, we show the differences between the estimated basis vectors found using the basic and the proposed NMF models for \(\beta \) divergence by applying it to real purchasing data.     

A new constrained fixed-point algorithm for ordering independent components

Independent component analysis (ICA) aims to recover a set of unknown mutually independent components (ICs) from their observed mixtures without knowledge of the mixing coefficients. In the classical ICA model there exists ICs’ indeterminacy on permutation and dilation. Constrained ICA is one of methods for solving this problem through introducing constraints into the classical ICA model. In this paper we first present a new constrained ICA model which composed of three parts: a maximum likelihood criterion as an objective function, statistical measures as inequality constraints and the normalization of demixing matrix as equality constraints. Next, we incorporate the new fixed-point (newFP) algorithm into this constrained ICA model to construct a new constrained fixed-point algorithm. Computation simulations on synthesized signals and speech signals demonstrate that this combination both can eliminate ICs’ indeterminacy to a certain extent, and can provide better performance. Moreover, comparison results with the existing algorithm verify the efficiency of our new algorithm furthermore, and show that it is more simple to implement than the existing algorithm due to its advantage of not using the learning rate. Finally, this new algorithm is also applied for the real-world fetal ECG data, experiment results further indicate the efficiency of the new constrained fixed-point algorithm.     

DC-NMF: nonnegative matrix factorization based on divide-and-conquer for fast clustering and topic modeling

The importance of unsupervised clustering and topic modeling is well recognized with ever-increasing volumes of text data available from numerous sources. Nonnegative matrix factorization (NMF) has proven to be a successful method for cluster and topic discovery in unlabeled data sets. In this paper, we propose a fast algorithm for computing NMF using a divide-and-conquer strategy, called DC-NMF. Given an input matrix where the columns represent data items, we build a binary tree structure of the data items using a recently-proposed efficient algorithm for computing rank-2 NMF, and then gather information from the tree to initialize the rank-k NMF, which needs only a few iterations to reach a desired solution. We also investigate various criteria for selecting the node to split when growing the tree. We demonstrate the scalability of our algorithm for computing general rank-k NMF as well as its effectiveness in clustering and topic modeling for large-scale text data sets, by comparing it to other frequently utilized state-of-the-art algorithms. The value of the proposed approach lies in the highly efficient and accurate method for initializing rank-k NMF and the scalability achieved from the divide-and-conquer approach of the algorithm and properties of rank-2 NMF. In summary, we present efficient tools for analyzing large-scale data sets, and techniques that can be generalized to many other data analytics problem domains along with an open-source software library called SmallK.     

Sequence distance via parsing complexity: Heartbeat signals

We compare and discuss the use of different symbolic codings of electrocardiogram (ECG) signals in order to distinguish healthy patients from hospitalized ones. To this aim, we recall a parsing-based similarity distance and compare the performances of several methods of classification of data.     

A new nonlinear blind source separation method with chaos indicators for decoupling diagnosis of hybrid failures: A marine propulsion gearbox case with a large speed variation

The normal operation of propulsion gearboxes ensures the ship safety. Chaos indicators could efficiently indicate the state change of the gearboxes. However, accurate detection of gearbox hybrid faults using Chaos indicators is a challenging task and the detection under speed variation conditions is attracting considerable attentions. Literature review suggests that the gearbox vibration is a kind of nonlinear mixture of variant vibration sources and the blind source separation (BSS) is reported to be a promising technique for fault vibration analysis, but very limited work has addressed the nonlinear BSS approach for hybrid faults decoupling diagnosis. Aiming to enhance the fault detection performance of Chaos indicators, this work presents a new nonlinear BSS algorithm for gearbox hybrid faults detection under a speed variation condition. This new method appropriately introduces the kernel spectral regression (KSR) framework into the morphological component analysis (MCA). The original vibration data are projected into the reproducing kernel Hilbert space (RKHS) where the instinct nonlinear structure in the original data can be linearized by KSR. Thus the MCA is able to deal with nonlinear BSS in the KSR space. Reliable hybrid faults decoupling is then achieved by this new nonlinear MCA (NMCA). Subsequently, by calculating the Chaos indicators of the decoupled fault components and comparing them with benchmarks, the hybrid faults can be precisely identified. Two specially designed case studies were implemented to evaluate the proposed NMCA-Chaos method on hybrid gear faults decoupling diagnosis. The performance of the NMCA-Chaos was compared with state of art techniques. The analysis results show high performance of the proposed method on hybrid faults detection in a marine propulsion gearbox with large speed variations.     

A new method for chaos control in communication systems

With the increasing needs of global communication, the improvement of secure communication is of vital importance. This study proposes a new scheme for establishing secure communication systems. The new scheme separates white Gaussian noises from the chaotic signals with modified Independent Component Analysis (ICA) and then controls each chaotic signal. This scheme is able to deal with white Gaussian noises in the natural world. However, the signals separated by traditional ICA shows opposite phase and unequal amplitude, making chaos control impossible. Our study proposed a modified ICA, which can calculate accurately the phase and amplitude and ensure control of the chaotic systems. The result indicates that our proposed system can successfully separate white Gaussian noise and stabilize all the chaotic signals.     

Nonuniqueness set and the problem of time-optimal motions in a velocity field

We study the problem of time-optimal transfer to the origin for motions described by the law $\dot x = v(x) + u(t)$ , where v: ?² → ?² is a smooth vector field and the control u satisfies the inequality ‖u(t)‖ ≤ u ₀. We introduce the notion of nonuniqueness set for problems with continuous optimal controls. By using the Pontryagin maximum principle, we single out a family of trajectories that lead to the origin and may be optimal. The nonuniqueness set intersects this family and cuts away the nonoptimal part from each trajectory. We show that the nonuniqueness set for a plane-parallel velocity field with symmetric profile is a ray lying on the symmetry axis. In the case of a nonsymmetric profile, we construct a Cauchy problem whose trajectory is the nonuniqueness set.     

Skew-prime polynomial matrices: The polynomial-model approach

We examine the concept of skew-primeness of polynomial matrices in terms of the associated polynomial model. It is shown that skew-primeness can be characterized in terms of the property of decomposition of a vector space relative to an endomorphism. This basic result is then applied to the special case of nonsingular polynomial matrices. We investigate the nonuniqueness of skew-complements of a skew-prime pair. It is shown that the space of equivalence classes of skew-complements is in bijective correspondence with a finite-dimensional linear space. Finally, the equivalence of the solutions to the problem of output regulation with internal stability obtained via geometric methods and via polynomial matrix techniques is shown explicitly.     

城市居民生活垃圾源头分类行为的影响因素研究

识别影响我国城市居民生活垃圾源头分类行为的影响因素是确保生活垃圾源头分类有效实施的基础和前提。本文以计划行为理论为基础,运用实证研究方法,分析了影响我国城市居民生活垃圾源头分类行为的影响因素,研究结果表明源头分类行为意向包含2个维度,影响因素包含7个维度,且借鉴层次分析方法解释说明了影响因素对源头分类行为的影响力。研究结论可以为城市市政相关部门制定政策和促进居民实施源头分类提供指导和借鉴。     

Semi-blind source extraction algorithm for fetal electrocardiogram based on generalized autocorrelations and reference signals

Blind source extraction (BSE) has become one of the promising methods in the field of signal processing and analysis, which only desires to extract “interesting” source signals with specific stochastic property or features so as to save lots of computing time and resources. This paper addresses BSE problem, in which desired source signals have some available reference signals. Based on this prior information, we develop an objective function for extraction of temporally correlated sources. Maximizing this objective function, a semi-blind source extraction fixed-point algorithm is proposed. Simulations on artificial electrocardiograph (ECG) signals and the real-world ECG data demonstrate the better performance of the new algorithm. Moreover, comparisons with existing algorithms further indicate the validity of our new algorithm, and also show its robustness to the estimated error of time delay.     

Convergence analysis of a non-negative matrix factorization algorithm based on Gibbs random field modeling

Non-negative matrix factorization (NMF) is a new approach to deal with the multivariate nonnegative data. Although the classic multiplicative update algorithm can solve the NMF problems, it fails to find sparse and localized object parts. Then a Gibbs random field (GRF) modeling based NMF algorithm, called the GRF-NMF algorithm, try to directly model the prior object structure of the components into the NMF problem. In this paper, the convergence of the GRF-NMF algorithm and its advantages are investigated. Based on a classic model, the equilibrium points are obtained. Some invariant sets are constructed to prepare for the analysis of the convergence of the GRF-NMF algorithm. Then using stability theory of the equilibrium point, the convergence of the algorithm is proved and the convergence conditions of the algorithm are obtained. We theoretically present the advantages of the GRF-NMF algorithm in the end.     

Modified subspace Barzilai-Borwein gradient method for non-negative matrix factorization

Non-negative matrix factorization (NMF) is a problem to obtain a representation of data using non-negativity constraints. Since the NMF was first proposed by Lee, NMF has attracted much attention for over a decade and has been successfully applied to numerous data analysis problems. Recent years, many variants of NMF have been proposed. Common methods are: iterative multiplicative update algorithms, gradient descent methods, alternating least squares (ANLS). Since alternating least squares has nice optimization properties, various optimization methods can be used to solve ANLS’s subproblems. In this paper, we propose a modified subspace Barzilai-Borwein for subproblems of ANLS. Moreover, we propose a modified strategy for ANLS. Global convergence results of our algorithm are established. The results of numerical experiments are reported to show the effectiveness of the proposed algorithm.     

基于独立分量分析——最小二乘支持向量机的公路建设项目投资估算新模型

为快速、准确地进行公路建设项目投资估算,提出了一种新型的公路建设项目投资估算模型.该模型首先基于独立分量分析技术,根据最小互信息原理,有效分离出公路建设项目投资估算的独立影响因素源.然后,将这些独立影响因素源用于最小二乘支持向量机的训练,从而建立了基于独立分量分析技术—最小二乘支持向量机的公路建设项目投资估算模型.该模型将独立分量分析技术的盲信号分离能力与最小二乘支持向量机处理有限样本条件下非线性回归问题的优势有机结合,提高了模型预测的准确性.     

ENHANCED BLOCK-SPARSE SIGNAL RECOVERY PERFORMANCE VIA TRUNCATED e2/e1-2 MINIMIZATION

In this paper, we investigate truncated $ℓ_2/ℓ_{1−2}$ minimization and its associated alternating direction method of multipliers (ADMM) algorithm for recovering the block sparse signals. Based on the block restricted isometry property (Block-RIP), a theoretical analysis is presented to guarantee the validity of proposed method. Our theoretical results not only show a less error upper bound, but also promote the former recovery condition of truncated ℓ₁₋₂ method for sparse signal recovery. Besides, the algorithm has been compared with some state-of-the-art algorithms and numerical experiments have shown excellent performances on recovering the block sparse signals.     

A note on semigroup arguments in nonstandard problems

In this note we investigate a nonstandard problem determined by a system of differential equations and initial conditions. We use the techniques of the mild solutions to obtain continuous dependence estimates, uniqueness and existence of solutions. We also make some comments on nonexistence and nonuniqueness of solutions.