Kernel PCA for Feature Extraction and De-Noising in Nonlinear Regression

In this paper, we propose the application of the Kernel Principal Component Analysis (PCA) technique for feature selection in a high-dimensional feature space, where input variables are mapped by a Gaussian kernel. The extracted features are employed in the regression problems of chaotic Mackey–Glass time-series prediction in a noisy environment and estimating human signal detection performance from brain event-related potentials elicited by task relevant signals. We compared results obtained using either Kernel PCA or linear PCA as data preprocessing steps. On the human signal detection task, we report the superiority of Kernel PCA feature extraction over linear PCA. Similar to linear PCA, we demonstrate de-noising of the original data by the appropriate selection of various nonlinear principal components. The theoretical relation and experimental comparison of Kernel Principal Components Regression, Kernel Ridge Regression and ε-insensitive Support Vector Regression is also provided.     

On-line Algorithm for Blind Signal Extraction of Arbitrarily Distributed,but Temporally Correlated Sources Using Second Order Statistics

Most of the algorithms for blind separation/extraction and independent component analysis (ICA) can not separate mixtures of sources with extremely low kurtosis or colored Gaussian sources. Moreover, to separate mixtures of super- and sub-Gaussian signals, it is necessary to use adaptive (time-variable) or switching nonlinearities which are controlled via computationally intensive measures, such as estimation of the sign of kurtosis of extracted signals. In this paper, we develop a very simple neural network model and an efficient on-line adaptive algorithm that sequentially extract temporally correlated sources with arbitrary distributions, including colored Gaussian sources and sources with extremely low values (or even zero) of kurtosis. The validity and performance of the algorithm have been confirmed by extensive computer simulation experiments.     

Three tokens in Herman’s algorithm

Herman’s algorithm is a synchronous randomized protocol for achieving self-stabilization in a token ring consisting of N processes. The interaction of tokens makes the dynamics of the protocol very difficult to analyze. In this paper we study the distribution of the time to stabilization, assuming that there are three tokens in the initial configuration. We show for arbitrary N and for an arbitrary timeout t that the probability of stabilization within time t is minimized by choosing as the initial three-token configuration the configuration in which the tokens are placed equidistantly on the ring. Our result strengthens a corollary of a theorem of McIver and Morgan (Inf. Process Lett. 94(2): 79–84, 2005), which states that the expected stabilization time is minimized by the equidistant configuration.     

Plasma lipids are affected similarly by dietary lauric or palmitic acid in gerbils and monkeys

To compare the relative impact of dietary lauric acid (12∶0) and palmitic acid (16∶0) on plasma lipids, two fat-sensitive species, Mongolian gerbils and cebus monkeys, were fed cholesterol-free, purified diets enriched with either 12∶0-rich or 16∶0-rich fats, while all other fatty acids were held constant by selective blending of up to five natural fats or oils. The two gerbil diets (40 en% from fat) allowed for an 8 en% exchange between 12∶0 and 16∶0, and the monkey diets (31 en% from fat) allowed for 6 en% exchange beteen these two fatty acids. Eight gerbils received the diets for eight weeks, and 12 cebus monkeys were fed each diet in a cross-over design for up to 22 wk. Both diets resulted in similar plasma cholesterol, triglyceride, and high density lipoprotein cholesterol concentrations within each species. Additionally, separation of cebus lipoproteins by discontinuous density-gradient ultracentrifugation failed to show any dietary differences in concentration or composition of the three major lipoprotein classes (d<1.019, 1.019–1.055, and 1.055–1.168 g/mL). Thus, in two species sensitive to manipulations in dietary fat while consuming cholesterol-free diets, 16∶0 was not hypercholesterolemic relative to 12∶0. Based on a paper presented at the PORIM International Palm Oil Congress (PIPOC) held in Kuala Lumpur, Malaysia, September 1993.     

Analysis of fluid flow and energy transport in Czochralski's process

In this paper the conjugate fluid flow and energy transport problem (involving conduction-convection-radiation heat transfer) resulting from the Czochralski crystal growth process is analysed. The solidifying material is treated as a pure and semitransparent substance with material properties depending neither on temperature nor on the wavelength. The solution of the problem is obtained iteratively using two computer codes: FLUENT, a commercial CFD package, and BEM-based in-house code capable of analysing the radiative heat transfer in the entire computational domain.Obtained results not only show velocity field and temperature distribution within the bodies under consideration but also demonstrate the influence of thermal radiation on these quantities.     

A Spurious Equilibria-free Learning Algorithm for the Blind Separation of Non-zoer Skewness Signals

We present a new learning algorithm for the blind separation of independent source signals having non-zero skewness (the 3rd-order cumulant) (the source signals have non-symmetric probability distribution.), from their linear mixtures. It is shown that for a class of source signals whose probability distribution functions is not symmetric, a simple adaptive learning algorithm using quadratic function (f(x)=x²) is very efficient for blind source separation task. It is proved that all stable equilibria of the proposed learning algorithm are desirable solutions. Extensive computer simulation experiments confirmed the validity of the proposed algorithm.     

Rough Sets in Reasoning about Vague Concepts

We discuss the rough set approach to some challenging problems that are of great importance for making progress in many applications. Among such challenges are the following ones.     

Methods of Image Compression by PHL Transform

In this paper an image data compression scheme based on Periodic Haar Piecewise-Linear (PHL) transform and quantization tables is proposed. Effectiveness of the compression for different classes of images is evaluated. The comparison of the compression quality using PHL and DCT transforms is given.     

Modification of Dugdale model to include the work hardening and in- and out-of-plane constraints

In the paper the classical Dugdale model has been generalized taking into account the influence of the specimen thickness, in-plane constraint as well as the effect of the strain hardening on the level of stress distribution within the strip yield zone (SYZ). Modification has been performed utilizing Huber, Mises, Hencky as well as Tresca yield hypotheses and Guo Wanlin Tz coefficient. Results are presented in a form useful for applications. As an example, the modified model has been applied to draw the failure assessment diagram (FAD). New FAD’s have been compared with others adopted from the SINTAP procedures.     

Formulation of Presliding Domain Non-local Memory Hysteretic Loops Based Upon Modified Maxwell Slip Model

In this paper, we investigate formulation of non-local memory effect hysteretic loops in presliding domain of friction by examining the dynamic responses of a single-degree-of-freedom friction oscillator. We modify the Maxwell slip model introducing characteristic features of asperities and perform detailed examination of the system from two perspectives. The first one, termed by us as microscale, presents contribution of each individual asperity to the total friction force, whereas the second one, termed by us as macroscale, concerns friction force behaviour of the whole model treated as the unity, with a special emphasis on non-local memory effect hysteretic loops formulation, complemented by time responses.