Automatic blood glucose classification for gestational diabetes with feature selection: decision trees vs neural networks

Automatic blood glucose classification may help specialists to provide a better interpretation of blood glucose data, downloaded directly from patients glucose meter and will contribute in the development of decision support systems for gestational diabetes. This paper presents an automatic blood glucose classifier for gestational diabetes that compares 6 different feature selection methods for two machine learning algorithms: neural networks and decision trees. Three searching algorithms, Greedy, Best First and Genetic, were combined with two different evaluators, CSF and Wrapper, for the feature selection. The study has been made with 6080 blood glucose measurements from 25 patients. Decision trees with a feature set selected with the Wrapper evaluator and the Best first search algorithm obtained the best accuracy: 95.92%.

A multi-resource load balancing algorithm for cloud cache systems

With the advent of cloud computing model, distributed caches have become the cornerstone for building scalable applications. Popular systems like Facebook [1] or Twitter use Memcached [5], a highly scalable distributed object cache, to speed up applications by avoiding database accesses. Distributed object caches assign objects to cache instances based on a hashing function, and objects are not moved from a cache instance to another unless more instances are added to the cache and objects are redistributed. This may lead to situations where some cache instances are overloaded when some of the objects they store are frequently accessed, while other cache instances are less frequently used. In this paper we propose a multi-resource load balancing algorithm for distributed cache systems. The algorithm aims at balancing both CPU and Memory resources among cache instances by redistributing stored data. Considering the possible conflict of balancing multiple resources at the same time, we give CPU and Memory resources weighted priorities based on the runtime load distributions. A scarcer resource is given a higher weight than a less scarce resource when load balancing. The system imbalance degree is evaluated based on monitoring information, and the utility load of a node, a unit for resource consumption. Besides, since continuous rebalance of the system may affect the QoS of applications utilizing the cache system, our data selection policy ensures that each data migration minimizes the system imbalance degree and hence, the total reconfiguration cost can be minimized. An extensive simulation is conducted to compare our policy with other policies. Our policy shows a significant improvement in time efficiency and decrease in reconfiguration cost.

Consideration of Multi-Phase Criterion in the Differential Protection Algorithm for High-Impedance Grounded Synchronous Generators

The generator differential protection is one of the most important electrical protections of synchronous generator stator windings. Its operation principle is based on the comparison of the input current and output current at each phase winding. Unwanted trip commands are usually caused by CT saturation, wrong CT selection, or the fact that they may come from different manufacturers. In generators grounded through high impedance, only phase-to-phase or three-phase faults can be detected by the differential protection. This kind of fault causes differential current to flow in, at least, two phases of the winding. Several cases of unwanted trip commands caused by the appearance of differential current in only one phase of the generator have been reported. In this paper multi-phase criterion is proposed for generator differential protection algorithm when applied to high impedance grounded generators.

Large aperture antenna arrays with subband delay compensation for reception of wideband satellite signals

Con esta disertación se pretenden resolver algunos de los problemas encontrados actualmente en la recepción de señales de satélites bajo dos escenarios particularmente exigentes: comunicaciones de Espacio Profundo y en banda Ka. Las comunicaciones con sondas de Espacio Profundo necesitan grandes aperturas en tierra para poder incrementar la velocidad de datos. La opción de usar antennas con diámetro mayor de 35 metros tiene serios problemas, pues antenas tan grandes son caras de mantener, difíciles de apuntar, pueden tener largos tiempo de reparación y además tienen una efeciencia decreciente a medida que se utilizan bandas más altas. Soluciones basadas en agrupaciones de antenas de menor tamaño (12 ó 35 metros) son mas ecónomicas y factibles técnicamente. Las comunicaciones en banda Ka tambien pueden beneficiarse de la combinación de múltiples antennas. Las antenas de menor tamaño son más fáciles de apuntar y además tienen un campo de visión mayor. Además, las técnicas de diversidad espacial pueden ser reemplazadas por una combinación de antenas para así incrementar el margen del enlace. La combinación de antenas muy alejadas sobre grandes anchos de banda, bien por recibir una señal de banda ancha o múltiples de banda estrecha, es complicada técnicamente. En esta disertación se demostrará que el uso de conformador de haz en el dominio de la frecuencia puede ayudar a relajar los requisitos de calibración y, al mismo tiempo, proporcionar un mayor campo de visión y mayores capacidades de ecualización. Para llevar esto a cabo, el trabajo ha girado en torno a tres aspectos fundamentales. El primero es la investigación bibliográfica del trabajo existente en este campo. El segundo es el modelado matemático del proceso de combinación y el desarrollo de nuevos algoritmos de estimación de fase y retardo. Y el tercero es la propuesta de nuevas aplicaciones en las que usar estas técnicas. La investigación bibliográfica se centra principalmente en los capítulos 1, 2, 4 y 5. El capítulo 1 da una breve introducción a la teoría de combinación de antenas de gran apertura. En este capítulo, los principales campos de aplicación son descritos y además se establece la necesidad de compensar retardos en subbandas. La teoría de bancos de filtros se expone en el capítulo 2; se selecciona y simula un banco de filtros modulado uniformemente con fase lineal. Las propiedades de convergencia de varios filtros adaptativos se muestran en el capítulo 4. Y finalmente, las técnicas de estimación de retardo son estudiadas y resumidas en el capítulo 5. Desde el punto de vista matemático, las principales contribución de esta disertación han sido: • Sección 3.1.4. Cálculo de la desviación de haz de un conformador de haz con compensación de retardo en pasos discretos en frecuencia intermedia. • Sección 3.2. Modelo matemático de un conformador de haz en subbandas. • Sección 3.2.2. Cálculo de la desviación de haz de un conformador de haz en subbandas con un buffer de retardo grueso. • Sección 3.2.4. Análisis de la influencia de los alias internos en la compensación en subbandas de retardo y fase. • Sección 3.2.4.2. Cálculo de la desviación de haz de un conformador de haz con compensación de retardo en subbandas. • Sección 3.2.6. Cálculo de la ganancia de relación señal a ruido de la agrupación de antenas en cada una de las subbandas. • Sección 3.3.2. Modelado de la función de transferencia de la agrupación de antenas bajo errores de estimación de retardo. • Sección 3.3.3. Modelado de los efectos de derivas de fase y retardo entre actualizaciones de las estimaciones. • Sección 3.4. Cálculo de la directividad de la agrupación de antenas con y sin compensación de retardos en subbandas. • Sección 5.2.6. Desarrollo de un algorimo para estimar la fase y el retardo entre dos señales a partir de su descomposición de subbandas bajo entornos estacionarios. • Sección 5.5.1. Desarrollo de un algorimo para estimar la fase, el retardo y la deriva de retardo entre dos señales a partir de su descomposición de subbandas bajo entornos no estacionarios. Las aplicaciones que se pueden beneficiar de estas técnicas son descritas en el capítulo 7: • Sección 6.2. Agrupaciones de antenas para comunicaciones de Espacio Profundo con capacidad multihaz y sin requisitos de calibración geométrica o de retardo de grupo. • Sección 6.2.6. Combinación en banda ancha de antenas con separaciones de miles de kilómetros, para recepción de sondas de espacio profundo. • Secciones 6.4 and 6.3. Combinación de estaciones remotas en banda Ka en escenarios de diversidad espacial, para recepción de satélites LEO o GEO. • Sección 6.3. Recepción de satélites GEO colocados con arrays de antenas multihaz. Las publicaciones a las que ha dado lugar esta tesis son las siguientes • A. Torre. Wideband antenna arraying over long distances. Interplanetary Progress Report, 42-194:1–18, 2013. En esta pulicación se resumen los resultados de las secciones 3.2, 3.2.2, 3.3.2, los algoritmos en las secciones 5.2.6, 5.5.1 y la aplicación destacada en 6.2.6. • A. Torre. Reception of wideband signals from geostationary collocated satellites with antenna arrays. IET Communications, Vol. 8, Issue 13:2229–2237, September, 2014. En esta segunda se muestran los resultados de la sección 3.2.4, el algoritmo en la sección 5.2.6.1 , y la aplicación mostrada en 6.3. ABSTRACT This dissertation is an attempt to solve some of the problems found nowadays in the reception of satellite signals under two particular challenging scenarios: Deep Space and Ka-band communications. Deep Space communications require from larger apertures on ground in order to increase the data rate. The option of using single dishes with diameters larger than 35 meters has severe drawbacks. Such antennas are expensive to maintain, prone to long downtimes, difficult to point and have a degraded performance in high frequency bands. The array solution, either with 12 meter or 35 meter antennas is deemed to be the most economically and technically feasible solution. Ka-band communications can also benefit from antenna arraying technology. The smaller aperture antennas that make up the array are easier to point and have a wider field of view allowing multiple simultaneous beams. Besides, site diversity techniques can be replaced by pure combination in order to increase link margin. Combination of far away antennas over a large bandwidth, either because a wideband signal or multiple narrowband signals are received, is a demanding task. This dissertation will show that the use of frequency domain beamformers with subband delay compensation can help to ease calibration requirements and, at the same time, provide with a wider field of view and enhanced equalization capabilities. In order to do so, the work has been focused on three main aspects. The first one is the bibliographic research of previous work on this subject. The second one is the mathematical modeling of the array combination process and the development of new phase/delay estimation algorithms. And the third one is the proposal of new applications in which these techniques can be used. Bibliographic research is mainly done in chapters 1, 2, 4 and 5. Chapter 1 gives a brief introduction to previous work in the field of large aperture antenna arraying. In this chapter, the main fields of application are described and the need for subband delay compensation is established. Filter bank theory is shown in chapter 2; a linear phase uniform modulated filter bank is selected and simulated under diverse conditions. The convergence properties of several adaptive filters are shown in chapter 4. Finally, delay estimation techniques are studied and summarized in chapter 5. From a mathematical point of view, the main contributions of this dissertation have been: • Section 3.1.4. Calculation of beam squint of an IF beamformer with delay compensation at discrete time steps. • Section 3.2. Establishment of a mathematical model of a subband beamformer. • Section 3.2.2. Calculation of beam squint in a subband beamformer with a coarse delay buffer. • Section 3.2.4. Analysis of the influence of internal aliasing on phase and delay subband compensation. • Section 3.2.4.2. Calculation of beam squint of a beamformer with subband delay compensation. • Section 3.2.6. Calculation of the array SNR gain at each of the subbands. • Section 3.3.2. Modeling of the transfer function of an array subject to delay estimation errors. • Section 3.3.3. Modeling of the effects of phase and delay drifts between estimation updates. • Section 3.4. Calculation of array directivity with and without subband delay compensation. • Section 5.2.6. Development of an algorithm to estimate relative delay and phase between two signals from their subband decomposition in stationary environments. • Section 5.5.1. Development of an algorithm to estimate relative delay rate, delay and phase between two signals from their subband decomposition in non stationary environments. The applications that can benefit from these techniques are described in chapter 7: • Section 6.2. Arrays of antennas for Deep Space communications with multibeam capacity and without geometric or group delay calibration requirement. • Section 6.2.6. Wideband antenna arraying over long distances, in the range of thousands of kilometers, for reception of Deep Space probes. • Sections 6.4 y 6.3. Combination of remote stations in Ka-band site diversity scenarios for reception of LEO or GEO satellites. • Section 6.3. Reception of GEO collocated satellites with multibeam antenna arrays. The publications that have been made from the work in this dissertation are • A. Torre. Wideband antenna arraying over long distances. Interplanetary Progress Report, 42-194:1–18, 2013. This article shows the results in sections 3.2, 3.2.2, 3.3.2, the algorithms in sections 5.2.6, 5.5.1 and the application in section 6.2.6. • A. Torre. Reception of wideband signals from geostationary collocated satellites with antenna arrays. IET Communications, Vol. 8, Issue 13:2229–2237, September, 2014. This second article shows among others the results in section 3.2.4, the algorithm in section 5.2.6.1 , and the application in section 6.3.

Improving parameters selection of a seeded region growing method for multiband image segmentation

In the last decade, Object Based Image Analysis (OBIA) has been accepted as an effective method for processing high spatial resolution multiband images. This image analysis method is an approach that starts with the segmentation of the image. Image segmentation in general is a procedure to partition an image into homogenous groups (segments). In practice, visual interpretation is often used to assess the quality of segmentation and the analysis relies on the experience of an analyst. In an effort to address the issue, in this study, we evaluate several seed selection strategies for an automatic image segmentation methodology based on a seeded region growing-merging approach. In order to evaluate the segmentation quality, segments were subjected to spatial autocorrelation analysis using Moran's I index and intra-segment variance analysis. We apply the algorithm to image segmentation using an aerial multiband image.

Improved damage detectability in a wind turbine blade by optimal selection of vibration signal correlation coefficients

Piotr Omenzetter and Simon Hoell’s work within the Lloyd’s Register Foundation Centre for Safety and Reliability Engineering at the University of Aberdeen is supported by Lloyd’s Register Foundation. The Foundation helps to protect life and property by supporting engineering-related education, public engagement and the application of research.

Three-dimensional inverse modelling of magnetic anomaly sources based on a genetic algorithm.

We present a modelling method to estimate the 3-D geometry and location of homogeneously magnetized sources from magnetic anomaly data. As input information, the procedure needs the parameters defining the magnetization vector (intensity, inclination and declination) and the Earth's magnetic field direction. When these two vectors are expected to be different in direction, we propose to estimate the magnetization direction from the magnetic map. Then, using this information, we apply an inversion approach based on a genetic algorithm which finds the geometry of the sources by seeking the optimum solution from an initial population of models in successive iterations through an evolutionary process. The evolution consists of three genetic operators (selection, crossover and mutation), which act on each generation, and a smoothing operator, which looks for the best fit to the observed data and a solution consisting of plausible compact sources. The method allows the use of non-gridded, non-planar and inaccurate anomaly data and non-regular subsurface partitions. In addition, neither constraints for the depth to the top of the sources nor an initial model are necessary, although previous models can be incorporated into the process. We show the results of a test using two complex synthetic anomalies to demonstrate the efficiency of our inversion method. The application to real data is illustrated with aeromagnetic data of the volcanic island of Gran Canaria (Canary Islands).

A novel information theory method for filter feature selection

In this paper, we propose a novel filter for feature selection. Such filter relies on the estimation of the mutual information between features and classes. We bypass the estimation of the probability density function with the aid of the entropic-graphs approximation of Rényi entropy, and the subsequent approximation of the Shannon one. The complexity of such bypassing process does not depend on the number of dimensions but on the number of patterns/samples, and thus the curse of dimensionality is circumvented. We show that it is then possible to outperform a greedy algorithm based on the maximal relevance and minimal redundancy criterion. We successfully test our method both in the contexts of image classification and microarray data classification.

Logic hybrid simulation-optimization algorithm for distillation design

In this paper, we propose a novel algorithm for the rigorous design of distillation columns that integrates a process simulator in a generalized disjunctive programming formulation. The optimal distillation column, or column sequence, is obtained by selecting, for each column section, among a set of column sections with different number of theoretical trays. The selection of thermodynamic models, properties estimation etc., are all in the simulation environment. All the numerical issues related to the convergence of distillation columns (or column sections) are also maintained in the simulation environment. The model is formulated as a Generalized Disjunctive Programming (GDP) problem and solved using the logic based outer approximation algorithm without MINLP reformulation. Some examples involving from a single column to thermally coupled sequence or extractive distillation shows the performance of the new algorithm.

Improving kNN multi-label classification in Prototype Selection scenarios using class proposals

Prototype Selection (PS) algorithms allow a faster Nearest Neighbor classification by keeping only the most profitable prototypes of the training set. In turn, these schemes typically lower the performance accuracy. In this work a new strategy for multi-label classifications tasks is proposed to solve this accuracy drop without the need of using all the training set. For that, given a new instance, the PS algorithm is used as a fast recommender system which retrieves the most likely classes. Then, the actual classification is performed only considering the prototypes from the initial training set belonging to the suggested classes. Results show that this strategy provides a large set of trade-off solutions which fills the gap between PS-based classification efficiency and conventional kNN accuracy. Furthermore, this scheme is not only able to, at best, reach the performance of conventional kNN with barely a third of distances computed, but it does also outperform the latter in noisy scenarios, proving to be a much more robust approach.

A genetic estimation algorithm for parameters of stochastic ordinary differential equations

A generic method for the estimation of parameters for Stochastic Ordinary Differential Equations (SODEs) is introduced and developed. This algorithm, called the GePERs method, utilises a genetic optimisation algorithm to minimise a stochastic objective function based on the Kolmogorov-Smirnov statistic. Numerical simulations are utilised to form the KS statistic. Further, the examination of some of the factors that improve the precision of the estimates is conducted. This method is used to estimate parameters of diffusion equations and jump-diffusion equations. It is also applied to the problem of model selection for the Queensland electricity market. (C) 2003 Elsevier B.V. All rights reserved.

Genetic parameters and predicted response to selection for lean growth in an Australian commercial nucleus pig herd

Genetic parameters for performance traits in a pig population were estimated using a multi-trait derivative-free REML algorithm. The 2590 total data included 922 restrictively fed male and 1668 ad libitum fed female records. Estimates of heritability (standard error in parentheses) were 0.25 (0.03), 0.15 (0.03), and 0.30 (0.05) for lifetime daily gain, test daily gain, and P2-fat depth in males, respectively; and 0.27 (0.04) and 0.38 (0.05) for average daily gain and P2-fat depth in females, respectively. The genetic correlation between P2-fat depth and test daily gain in males was -0.17 (0.06) and between P2-fat and lifetime average daily gain in females 0.44 (0.09). Genetic correlations between sexes were 0.71 (0.11) for average daily gain and -0.30 (0.10) for P2-fat depth. Genetic response per standard deviation of selection on an index combining all traits was predicted at $AU120 per sow per year. Responses in daily gain and backfat were expected to be higher when using only male selection than when using only female selection. Selection for growth rate in males will improve growth rate and carcass leanness simultaneously.

The dynamics of a genetic algorithm for a simple learning problem

A formalism for describing the dynamics of Genetic Algorithms (GAs) using method s from statistical mechanics is applied to the problem of generalization in a perceptron with binary weights. The dynamics are solved for the case where a new batch of training patterns is presented to each population member each generation, which considerably simplifies the calculation. The theory is shown to agree closely to simulations of a real GA averaged over many runs, accurately predicting the mean best solution found. For weak selection and large problem size the difference equations describing the dynamics can be expressed analytically and we find that the effects of noise due to the finite size of each training batch can be removed by increasing the population size appropriately. If this population resizing is used, one can deduce the most computationally efficient size of training batch each generation. For independent patterns this choice also gives the minimum total number of training patterns used. Although using independent patterns is a very inefficient use of training patterns in general, this work may also prove useful for determining the optimum batch size in the case where patterns are recycled.

EM algorithm for GTM-FS

We propose a generative topographic mapping (GTM) based data visualization with simultaneous feature selection (GTM-FS) approach which not only provides a better visualization by modeling irrelevant features ("noise") using a separate shared distribution but also gives a saliency value for each feature which helps the user to assess their significance. This technical report presents a varient of the Expectation-Maximization (EM) algorithm for GTM-FS.

A constructive learning algorithm based on back-propagation

There are been a resurgence of interest in the neural networks field in recent years, provoked in part by the discovery of the properties of multi-layer networks. This interest has in turn raised questions about the possibility of making neural network behaviour more adaptive by automating some of the processes involved. Prior to these particular questions, the process of determining the parameters and network architecture required to solve a given problem had been a time consuming activity. A number of researchers have attempted to address these issues by automating these processes, concentrating in particular on the dynamic selection of an appropriate network architecture.The work presented here specifically explores the area of automatic architecture selection; it focuses upon the design and implementation of a dynamic algorithm based on the Back-Propagation learning algorithm. The algorithm constructs a single hidden layer as the learning process proceeds using individual pattern error as the basis of unit insertion. This algorithm is applied to several problems of differing type and complexity and is found to produce near minimal architectures that are shown to have a high level of generalisation ability.