Search

Blind source separation (BSS) involves recovering unobserved source signals from several mixed observations, typically obtained at the output of a set of sensors. Each sensor receives a different combination of the source signals. The adjective “blind” emphasizes the fact that: first, the source signals are not observed; and next, no information is available about the mixture. The assumption is often held physically that the source signals are mutually independent. Recently, BSS in signal processing has received considerable attention from researchers, due to its numerous promising applications in the areas of biomedical signal processing, digital communications and speech signal, sonar, image processing, and monitoring (Cichocki & Unbehauen, 1996), (Tangdiongga et al, 2001), (Yilmaz & Rickard, 2004), (Herault & Juten, 1986). A number of blind separation algorithms have been proposed based on different separation models. These algorithms play increasingly important roles in many applications. Since the pioneering work of Jutten and Herault (Herault & Juten, 1986), a variety of algorithms have been proposed for BSS. In general, the existing algorithms can be divided into five major categories: neural network-based algorithms (Cichocki & Unbehauen, 1996), (Zhang & Kassam, 2004), density model-based algorithms (Amari et al, 1997), (Lee et al, 1999a), algebraic algorithms (Belouchrani et al, 1997), (Li & Wang, 2002), information-theoretic algorithms (Pajunen, 1998), (Pham & Vrins, 2005) and space-based algorithms (Yilmaz & Rickard, 2004), (Lee et al, 1999b). A source signal with sparse representation means that at most one value of the signal isn’t zero at an instant, making the vector of sensor signals (mixtures) equivalent to some mixing vector. Therefore, the sparse-based BSS problem could be solved by searching for mixing vectors; moreover, recovering source signals. Like the mixtures with sparse representation, each base vectors of the unknown mixing matrix will be displayed on a 2-D plane coordinate as a directional line when two sensors are used. The sparse representation is first introduced in underdetermined BSS by Lee et al. (Lee et al, 1999b). After its introduction, several related methods have been proposed continuously for solving underdetermined BSS

This paper presents a macrocell placement constraint and overlap removal methodology using an improved particle swarm optimization (PSO). Several techniques have been proposed to improve PSO, such as methods to prevent the floorplan from falling into the local minimum and to assist in finding the global minimum. The proposed method can deal with various kinds of placement constraints and can process them simultaneously. Experiments employing MCNC and GSRC benchmarks show the difference in the efficiency and robustness of proposed method in the exploration for more optimal solutions through restricted placement and overlap removal compared with other methods. The proposed approach exhibits rapid convergence and leads to more optimal solutions than other related approaches; furthermore, it displays efficient packing with all the constraints satisfied.

In an attempt to build a good predictor of the performance of survey interviewers, we propose a feature selection method that derives the featuresi¦ strength (i.e., degree of usefulness) from various feature subsets drawn from a pool of all the features. The method also builds a predictor by using support vector regression (SVR) as the learning machine and the selected features as variables. Applying the method to a collection of 278 instances obtained from 67 interviewers par-ticipating in eight survey projects, we identified three critical features, experience and two attributional style variables, out of fifteen features. Compared with results of four existing methods, the proposed predictor produced the smallest predictive error. Furthermore, the three features utilized by our method were also identified as the most important features by the four compared methods.

To handle problems created by large data sets, we propose a method that uses a decision tree to decompose a given data space and trains SVMs on the decomposed regions. Although there are other means of decomposing a data space, we show that the decision tree has several merits for large-scale SVM training. First, it can classify some data points by its own means, thereby reducing the cost of SVM training applied to the remaining data points. Second, it is efficient for seeking the parameter values that maximize the validation accuracy, which helps maintain good test accuracy. For experiment data sets whose size can be handled by current non-linear, or kernel-based, SVM training techniques, the proposed method can speed up the training by a factor of thousands, and still achieve comparable test accuracy.

This paper addresses on recovering signals from white noise degradation without prior knowledge about pure signals. For the issue, wavelet-based methods support some acceptable solution, but their adaptability to various signal types is not good enough to perform more excellent. In order to deal the problem, this paper proposes an evolutionary strategy to choose a better split point of wavelet shrinkage. Moreover, the correlation evaluation of signals applied on evolutionary optimization makes the proposed method to remove noise more efficient. The advantage is demonstrated by simulation experiment having some benchmark cases and various noise degradations.

Noise reduction has a lot attention no matter in practical applications or a signal processing research field. Recently, a novel denoisy method which removes noise from received signals by threshold operation on wavelet coefficients was developed and its efficiency has been confirmed. However, its definition of parameters is not general-purpose enough to deal with variant cases. In order to seek high quality to denoisy, this study introduces a frequency analysis tool, empirical mode decomposition (EMD), to separate received signal into several elements which are termed intrinsic mode functions (IMF). And then, according to the order from high frequency to low frequency, IMFs could be separated into finite pair consisting of estimated noise and estimated original signal. Each estimated pair is calculated the correlation measurement which involves second-order correlation and high-order correlation since original signal and noise are mutually independent. A smallest measure value implies an optimal pair approximating to the real. In simulations, four benchmarks and three noise level are tested; moreover, two state of the art algorithms are compared with the proposed method. Finally, the excellent robustness and efficiency of proposed method are demonstrated by simulation results.

Radial Basis Function neural network (RBFNN) is a combination of learning vector quantizer LVQ-I and gradient descent. RBFNN is first proposed by (Broomhead & Lowe, 1988), and their interpolation and generalization properties are thoroughly investigated in (Lowe, 1989), (Freeman & Saad, 1995). Since the mid-1980s, RBFNN has been used to apply on many applications, such as pattern classification, system identification, nonlinear function approximation, adaptive control, speech recognition, and time-series prediction, and so on. In contrast to the well-known Multilayer Perceptron (MLP) Networks, the RBF network utilizes a radial construction mechanism. MLP were trained by the error Back Propagation (BP) algorithm, since the RBFNN has a faster training procedure substantially and adopts typical two-stage training scheme, it can avoid solution to fall into local optima. A key point of RBFNN is to decide a proper number of hidden nodes. If the hidden node number of RBFNN is too small, the generated output vectors may be in low accuracy. On the contrary, it with too large number of hidden nodes may cause over-fitting for the input data, and influences global generalization performance. In conventional RBF training approach, the number of hidden node is usually decided according to the statistic properties of input data, then determine the centers and spread width for each hidden nodes by means of k-means clustering algorithm (Moddy & Darken, 1989). The drawback of this approach is that the network performance is depended on the pre-selected number of hidden nodes. If an unsuitable number is chosen, RBFNN may present a poor global generalization capability, as slow training speed, and requirement for large memory space. To solve this problem, the self-growing RBF techniques were proposed in (Karayiannis & Mi, 1997), (Zheng et al, 1999). However, the predefined parameters and local searching on solution space cause the inaccuracy of approximation from a sub-solution. Evolutionary computation is a globally optimization technique, where the aim is to improve the ability of individual to survive. Among that, Genetic Algorithm (GA) is a parallel searching technique that mimics natural genetics and the evolutionary process. In (Back et al, 1997), they employed GA to determine the RBFNN structure so the optimal number and distribution of RBF hidden nodes can be obtained automatically. A common approach is applied GA to search for the optimal network structure among several candidates

For the research field of adaptive de-noisy, the technique which determinates an adequate threshold in wavelet domain of signal is novel and feasible. The most of threshold determination are developed from universal method proposed by Donoho. Unfortunately, these methods are just performed in some wavelet level and involve several incorrectly estimated factors; therefore, they can't result the best performance and can't work well in general cases. By the reason, this paper replaces the universal threshold determination by an intelligent determination, genetic algorithm (GA). Because original signals and noise are mutually independent, an objective function of GA is created to evaluate the second order correlation and high order correlation. Moreover, GA searching is applied in progressively wavelet levels to explore the most suitable wavelet decomposition and the optimal threshold. In order to confirm the validity and efficiency of the proposed algorithm, several simulations which include four benchmarks with variant noise degrees are designed. Moreover, the performance of proposed algorithm will have compared with that of other existing algorithms.

In this paper, a sharing evolution genetic algorithms (SEGA) is proposed to solve global numerical optimization problems. In the SEGA, three strategies are proposed, which are population manager, sharing cross-over and sharing mutation, for effective increasing new born offspring's solution searching ability. Experiments were conducted on CEC-05 benchmark problems which included unimodal, multimodal, expanded, and hybrid composition functions. The results showed that the proposed method exhibits better performance when solving these benchmark problems compared to recent variants of the genetic algorithms.

A very important step for the RBF network training is to decide a proper number of hidden nodes. This paper proposes a PSO-based algorithm for searching optimal cluster distance factor. Thus, the self-growing RBF network training algorithm can be realized by employing the optimal cluster distance factor to create hidden neurons automatically from the input data set. Two experiments that include function approximation and chaotic time series prediction are used to compare our proposed method with other related approaches. The PSO-based RBF network exhibits better generalization performance and shorter training time.

A self-organized rule is proposed to process the separation of unknown number of sources of blind signal. The rule is applied to the feed-forward neural network (FFNN) which is based on a mean weighting gradient (MWG) function. The algorithms adjust the architecture of neural network. For both, the performances are the mixture of node number and MWG threshold /spl xi/.The separating number of sources estimated accurately by experiment of computer simulations.

Blind source separation (BSS) is a technique for recovering a set of source signals without a priori information on the transformation matrix or the probability distributions of the source signals. In the previous works of BSS, the choice of the learning rate would reflect a trade-off between the stability and the speed of convergence. In this paper, we adapted the particle swarm optimization (PSO) technique to find suitable learning rates for each signal in each time slot. Experiments employing four mixed source signals were separated by our work and compared with other related approaches. The proposed approach exhibited rapid convergence and made the independent component analysis (ICA) algorithms become more efficient and stable than other related approaches.

Background Cancer therapy is a challenging research area because side effects often occur in chemo and radiation therapy. We intend to study a multi-targets and multi-components design that will provide synergistic results to improve efficiency of cancer therapy. Methods We have developed a general methodology, AMFES (Adaptive Multiple FEature Selection), for ranking and selecting important cancer biomarkers based on SVM (Support Vector Machine) classification. In particular, we exemplify this method by three datasets: a prostate cancer (three stages), a breast cancer (four subtypes), and another prostate cancer (normal vs. cancerous). Moreover, we have computed the target networks of these biomarkers as the signatures of the cancers with additional information (mutual information between biomarkers of the network). Then, we proposed a robust framework for synergistic therapy design approach which includes varies existing mechanisms. Results These methodologies were applied to three GEO datasets: GSE18655 (three prostate stages), GSE19536 (4 subtypes breast cancers) and GSE21036 (prostate cancer cells and normal cells) shown in. We selected 96 biomarkers for first prostate cancer dataset (three prostate stages), 72 for breast cancer (luminal A vs. luminal B), 68 for breast cancer (basal-like vs. normal-like), and 22 for another prostate cancer (cancerous vs. normal. In addition, we obtained statistically significant results of mutual information, which demonstrate that the dependencies among these biomarkers can be positive or negative. Conclusions We proposed an efficient feature ranking and selection scheme, AMFES, to select an important subset from a large number of features for any cancer dataset. Thus, we obtained the signatures of these cancers by building their target networks. Finally, we proposed a robust framework of synergistic therapy for cancer patients. Our framework is not only supported by real GEO datasets but also aim to a multi-targets/multi-components drug design tool, which improves the traditional single target/single component analysis methods. This framework builds a computational foundation which can provide a clear classification of cancers and lead to an efficient cancer therapy.