Exploration vs exploitation in differential evolution

Differential Evolution (DE) is a tool for efficient optimisation, and it belongs to the class of evolutionary algorithms, which include Evolution Strategies and Genetic Algorithms. DE algorithms work well when the population covers the entire search space, and they have shown to be effective on a large range of classical optimisation problems. However, an undesirable behaviour was detected when all the members of the population are in a basin of attraction of a local optimum (local minimum or local maximum), because in this situation the population cannot escape from it. This paper proposes a modification of the standard mechanisms in DE algorithm in order to change the exploration vs. exploitation balance to improve its behaviour.

Hammerstein model identification algorithm using Bezier-Bernstein approximation

A new identification algorithm is introduced for the Hammerstein model consisting of a nonlinear static function followed by a linear dynamical model. The nonlinear static function is characterised by using the Bezier-Bernstein approximation. The identification method is based on a hybrid scheme including the applications of the inverse of de Casteljau's algorithm, the least squares algorithm and the Gauss-Newton algorithm subject to constraints. The related work and the extension of the proposed algorithm to multi-input multi-output systems are discussed. Numerical examples including systems with some hard nonlinearities are used to illustrate the efficacy of the proposed approach through comparisons with other approaches.

Power properties if invariant tests for spatial autocorrelation in linear regression

This paper derives some exact power properties of tests for spatial autocorrelation in the context of a linear regression model. In particular, we characterize the circumstances in which the power vanishes as the autocorrelation increases, thus extending the work of Krämer (2005). More generally, the analysis in the paper sheds new light on how the power of tests for spatial autocorrelation is affected by the matrix of regressors and by the spatial structure. We mainly focus on the problem of residual spatial autocorrelation, in which case it is appropriate to restrict attention to the class of invariant tests, but we also consider the case when the autocorrelation is due to the presence of a spatially lagged dependent variable among the regressors. A numerical study aimed at assessing the practical relevance of the theoretical results is included

A mixture of experts network structure construction algorithm for modelling and control

This paper introduces a new fast, effective and practical model structure construction algorithm for a mixture of experts network system utilising only process data. The algorithm is based on a novel forward constrained regression procedure. Given a full set of the experts as potential model bases, the structure construction algorithm, formed on the forward constrained regression procedure, selects the most significant model base one by one so as to minimise the overall system approximation error at each iteration, while the gate parameters in the mixture of experts network system are accordingly adjusted so as to satisfy the convex constraints required in the derivation of the forward constrained regression procedure. The procedure continues until a proper system model is constructed that utilises some or all of the experts. A pruning algorithm of the consequent mixture of experts network system is also derived to generate an overall parsimonious construction algorithm. Numerical examples are provided to demonstrate the effectiveness of the new algorithms. The mixture of experts network framework can be applied to a wide variety of applications ranging from multiple model controller synthesis to multi-sensor data fusion.

PDM 2006: Proceedings of the International Workshop on Data Mining

Recently major processor manufacturers have announced a dramatic shift in their paradigm to increase computing power over the coming years. Instead of focusing on faster clock speeds and more powerful single core CPUs, the trend clearly goes towards multi core systems. This will also result in a paradigm shift for the development of algorithms for computationally expensive tasks, such as data mining applications. Obviously, work on parallel algorithms is not new per se but concentrated efforts in the many application domains are still missing. Multi-core systems, but also clusters of workstations and even large-scale distributed computing infrastructures provide new opportunities and pose new challenges for the design of parallel and distributed algorithms. Since data mining and machine learning systems rely on high performance computing systems, research on the corresponding algorithms must be on the forefront of parallel algorithm research in order to keep pushing data mining and machine learning applications to be more powerful and, especially for the former, interactive. To bring together researchers and practitioners working in this exciting field, a workshop on parallel data mining was organized as part of PKDD/ECML 2006 (Berlin, Germany). The six contributions selected for the program describe various aspects of data mining and machine learning approaches featuring low to high degrees of parallelism: The first contribution focuses the classic problem of distributed association rule mining and focuses on communication efficiency to improve the state of the art. After this a parallelization technique for speeding up decision tree construction by means of thread-level parallelism for shared memory systems is presented. The next paper discusses the design of a parallel approach for dis- tributed memory systems of the frequent subgraphs mining problem. This approach is based on a hierarchical communication topology to solve issues related to multi-domain computational envi- ronments. The forth paper describes the combined use and the customization of software packages to facilitate a top down parallelism in the tuning of Support Vector Machines (SVM) and the next contribution presents an interesting idea concerning parallel training of Conditional Random Fields (CRFs) and motivates their use in labeling sequential data. The last contribution finally focuses on very efficient feature selection. It describes a parallel algorithm for feature selection from random subsets. Selecting the papers included in this volume would not have been possible without the help of an international Program Committee that has provided detailed reviews for each paper. We would like to also thank Matthew Otey who helped with publicity for the workshop.

The CtsR regulator of Listeria monocytogenes contains a variant glycine repeat region that affects piezotolerance, stress resistance, motility and virulence

A spontaneous high hydrostatic pressure (HHP)-tolerant mutant of Listeria monocytogenes ScottA, named AK01, was isolated previously. This mutant was immotile and showed increased resistance to heat, acid and H2O2 compared with the wild type (wt) (Karatzas, K.A.G. and Bennik, M.H.J. 2002 Appl Environ Microbiol 68: 3183–3189). In this study, we conclusively linked the increased HHP and stress tolerance of strain AK01 to a single codon deletion in ctsR (class three stress gene repressor) in a region encoding a highly conserved glycine repeat. CtsR negatively regulates the expression of the clp genes, including clpP, clpE and the clpC operon (encompassing ctsR itself), which belong to the class III heat shock genes. Allelic replacement of the ctsR gene in the wt background with the mutant ctsR gene, designated ctsRΔGly, rendered mutants with phenotypes and protein expression profiles identical to those of strain AK01. The expression levels of CtsR, ClpC and ClpP proteins were significantly higher in ctsRΔGly mutants than in the wt strain, indicative of the CtsRΔGly protein being inactive. Further evidence that the CtsRΔGly protein lacks its repressor function came from the finding that the Clp proteins in the mutant were not further induced upon heat shock, and that HHP tolerance of a ctsR deletion strain was as high as that of a ctsRΔGly mutant. The high HHP tolerance possibly results from the increased expression of the clp genes in the absence of (active) CtsR repressor. Importantly, the strains expressing CtsRΔGly show significantly attenuated virulence compared with the wt strain; however, no indication of disregulation of PrfA in the mutant strains was found. Our data highlight an important regulatory role of the glycine-rich region of CtsR in stress resistance and virulence.

A conceptual framework for aligning business and information systems from organisational learning perspective

The three decades of on-going executives’ concerns of how to achieve successful alignment between business and information technology shows the complexity of such a vital process. Most of the challenges of alignment are related to knowledge and organisational change and several researchers have introduced a number of mechanisms to address some of these challenges. However, these mechanisms pay less attention to multi-level effects, which results in a limited un-derstanding of alignment across levels. Therefore, we reviewed these challenges from a multi-level learning perspective and found that business and IT alignment is related to the balance of exploitation and exploration strategies with the intellec-tual content of individual, group and organisational levels.

A unified neurofuzzy model for classification

This work proposes a unified neurofuzzy modelling scheme. To begin with, the initial fuzzy base construction method is based on fuzzy clustering utilising a Gaussian mixture model (GMM) combined with the analysis of covariance (ANOVA) decomposition in order to obtain more compact univariate and bivariate membership functions over the subspaces of the input features. The mean and covariance of the Gaussian membership functions are found by the expectation maximisation (EM) algorithm with the merit of revealing the underlying density distribution of system inputs. The resultant set of membership functions forms the basis of the generalised fuzzy model (GFM) inference engine. The model structure and parameters of this neurofuzzy model are identified via the supervised subspace orthogonal least square (OLS) learning. Finally, instead of providing deterministic class label as model output by convention, a logistic regression model is applied to present the classifier’s output, in which the sigmoid type of logistic transfer function scales the outputs of the neurofuzzy model to the class probability. Experimental validation results are presented to demonstrate the effectiveness of the proposed neurofuzzy modelling scheme.

Infra-red laser ablative micromachining of parylene-C on SiO2substrates for rapid prototyping, high yield, human neuronal cell patterning

Cell patterning commonly employs photolithographic methods for the micro fabrication of structures on silicon chips. These require expensive photo-mask development and complex photolithographic processing. Laser based patterning of cells has been studied in vitro and laser ablation of polymers is an active area of research promising high aspect ratios. This paper disseminates how 800 nm femtosecond infrared (IR) laser radiation can be successfully used to perform laser ablative micromachining of parylene-C on SiO2 substrates for the patterning of human hNT astrocytes (derived from the human teratocarcinoma cell line (hNT)) whilst 248 nm nanosecond ultra-violet laser radiation produces photo-oxidization of the parylene-C and destroys cell patterning. In this work, we report the laser ablation methods used and the ablation characteristics of parylene-C for IR pulse fluences. Results follow that support the validity of using IR laser ablative micromachining for patterning human hNT astrocytes cells. We disseminate the variation in yield of patterned hNT astrocytes on parylene-C with laser pulse spacing, pulse number, pulse fluence and parylene-C strip width. The findings demonstrate how laser ablative micromachining of parylene-C on SiO2 substrates can offer an accessible alternative for rapid prototyping, high yield cell patterning with broad application to multi-electrode arrays, cellular micro-arrays and microfluidics.

Guided growth of neurons and glia using microfabricated patterns of parylene-C on a SiO2 background

This paper describes a simple technique for the patterning of glia and neurons. The integration of neuronal patterning to Multi-Electrode Arrays (MEAs), planar patch clamp and silicon based ‘lab on a chip’ technologies necessitates the development of a microfabrication-compatible method, which will be reliable and easy to implement. In this study a highly consistent, straightforward and cost effective cell patterning scheme has been developed. It is based on two common ingredients: the polymer parylene-C and horse serum. Parylene-C is deposited and photo-lithographically patterned on silicon oxide (SiO2) surfaces. Subsequently, the patterns are activated via immersion in horse serum. Compared to non-activated controls, cells on the treated samples exhibited a significantly higher conformity to underlying parylene stripes. The immersion time of the patterns was reduced from 24 to 3 h without compromising the technique. X-ray photoelectron spectroscopy (XPS) analysis of parylene and SiO2 surfaces before and after immersion in horse serum and gel based eluant analysis suggests that the quantity and conformation of proteins on the parylene and SiO2 substrates might be responsible for inducing glial and neuronal patterning.

A novel transport mechanism for MOMP in Chlamydophila pneumoniae and its putative role in immune-therapy

Major outer membrane proteins (MOMPs) of Gram negative bacteria are one of the most intensively studied membrane proteins. MOMPs are essential for maintaining the structural integrity of bacterial outer membranes and in adaptation of parasites to their hosts. There is evidence to suggest a role for purified MOMP from Chlamydophila pneumoniae and corresponding MOMP-derived peptides in immune-modulation, leading to a reduced atherosclerotic phenotype in apoE−/− mice via a characteristic dampening of MHC class II activity. The work reported herein tests this hypothesis by employing a combination of homology modelling and docking to examine the detailed molecular interactions that may be responsible. A three-dimensional homology model of the C. pneumoniae MOMP was constructed based on the 14 transmembrane β-barrel crystal structure of the fatty acid transporter from Escherichia coli, which provides a plausible transport mechanism for MOMP. Ligand docking experiments were used to provide details of the possible molecular interactions driving the binding of MOMP-derived peptides to MHC class II alleles known to be strongly associated with inflammation. The docking experiments were corroborated by predictions from conventional immuno-informatic algorithms. This work supports further the use of MOMP in C. pneumoniae as a possible vaccine target and the role of MOMP-derived peptides as vaccine candidates for immune-therapy in chronic inflammation that can result in cardiovascular events.

Generation and analysis of the 2005 land cover map for Mexico using 250m MODIS data

Land cover plays a key role in global to regional monitoring and modeling because it affects and is being affected by climate change and thus became one of the essential variables for climate change studies. National and international organizations require timely and accurate land cover information for reporting and management actions. The North American Land Change Monitoring System (NALCMS) is an international cooperation of organizations and entities of Canada, the United States, and Mexico to map land cover change of North America's changing environment. This paper presents the methodology to derive the land cover map of Mexico for the year 2005 which was integrated in the NALCMS continental map. Based on a time series of 250 m Moderate Resolution Imaging Spectroradiometer (MODIS) data and an extensive sample data base the complexity of the Mexican landscape required a specific approach to reflect land cover heterogeneity. To estimate the proportion of each land cover class for every pixel several decision tree classifications were combined to obtain class membership maps which were finally converted to a discrete map accompanied by a confidence estimate. The map yielded an overall accuracy of 82.5% (Kappa of 0.79) for pixels with at least 50% map confidence (71.3% of the data). An additional assessment with 780 randomly stratified samples and primary and alternative calls in the reference data to account for ambiguity indicated 83.4% overall accuracy (Kappa of 0.80). A high agreement of 83.6% for all pixels and 92.6% for pixels with a map confidence of more than 50% was found for the comparison between the land cover maps of 2005 and 2006. Further wall-to-wall comparisons to related land cover maps resulted in 56.6% agreement with the MODIS land cover product and a congruence of 49.5 with Globcover.

The uniqueness of signature problem in the non-Markov setting

We establish a general framework for a class of multidimensional stochastic processes over [0,1] under which with probability one, the signature (the collection of iterated path integrals in the sense of rough paths) is well-defined and determines the sample paths of the process up to reparametrization. In particular, by using the Malliavin calculus we show that our method applies to a class of Gaussian processes including fractional Brownian motion with Hurst parameter H>1/4, the Ornstein–Uhlenbeck process and the Brownian bridge.

Atmospheric electrification in dusty, reactive gases in the solar system and beyond

Detailed observations of the solar system planets reveal a wide variety of local atmospheric conditions. Astronomical observations have revealed a variety of extrasolar planets none of which resembles any of the solar system planets in full. Instead, the most massive amongst the extrasolar planets, the gas giants, appear very similar to the class of (young) Brown Dwarfs which are amongst the oldest objects in the universe. Despite of this diversity, solar system planets, extrasolar planets and Brown Dwarfs have broadly similar global temperatures between 300K and 2500K. In consequence, clouds of different chemical species form in their atmospheres. While the details of these clouds differ, the fundamental physical processes are the same. Further to this, all these objects were observed to produce radio and X-ray emission. While both kinds of radiation are well studied on Earth and to a lesser extent on the solar system planets, the occurrence of emission that potentially originate from accelerated electrons on Brown Dwarfs, extrasolar planets and protoplanetary disks is not well understood yet. This paper offers an interdisciplinary view on electrification processes and their feedback on their hosting environment in meteorology, volcanology, planetology and research on extrasolar planets and planet formation.

An accelerating cosmology without dark energy

The negative pressure accompanying gravitationally-induced particle creation can lead to a cold dark matter (CDM) dominated, accelerating Universe (Lima et al. 1996 [1]) without requiring the presence of dark energy or a cosmological constant. In a recent study, Lima et al. 2008 [2] (LSS) demonstrated that particle creation driven cosmological models are capable of accounting for the SNIa observations [3] of the recent transition from a decelerating to an accelerating Universe, without the need for Dark Energy. Here we consider a class of such models where the particle creation rate is assumed to be of the form Gamma = beta H + gamma H(0), where H is the Hubble parameter and H(0) is its present value. The evolution of such models is tested at low redshift by the latest SNe Ia data provided by the Union compilation [4] and at high redshift using the value of z(eq), the redshift of the epoch of matter - radiation equality, inferred from the WMAP constraints on the early Integrated Sachs-Wolfe (ISW) effect [5]. Since the contributions of baryons and radiation were ignored in the work of LSS, we include them in our study of this class of models. The parameters of these more realistic models with continuous creation of CDM are constrained at widely-separated epochs (z(eq) approximate to 3000 and z approximate to 0) in the evolution of the Universe. The comparison of the parameter values, {beta, gamma}, determined at these different epochs reveals a tension between the values favored by the high redshift CMB constraint on z(eq) from the ISW and those which follow from the low redshift SNIa data, posing a potential challenge to this class of models. While for beta = 0 this conflict is only at less than or similar to 2 sigma, it worsens as beta increases from zero.