Comparing Support Vector Machines with Gaussian Kernels to Radial Basis Function Classifiers

The Support Vector (SV) machine is a novel type of learning machine, based on statistical learning theory, which contains polynomial classifiers, neural networks, and radial basis function (RBF) networks as special cases. In the RBF case, the SV algorithm automatically determines centers, weights and threshold such as to minimize an upper bound on the expected test error. The present study is devoted to an experimental comparison of these machines with a classical approach, where the centers are determined by $k$--means clustering and the weights are found using error backpropagation. We consider three machines, namely a classical RBF machine, an SV machine with Gaussian kernel, and a hybrid system with the centers determined by the SV method and the weights trained by error backpropagation. Our results show that on the US postal service database of handwritten digits, the SV machine achieves the highest test accuracy, followed by the hybrid approach. The SV approach is thus not only theoretically well--founded, but also superior in a practical application.

Percepciones del clima social escolar : diferencias tipológicas entre colegios públicos y colegios privados.

Identificar en dos muestras de población escolar urbana de Asturias, una perteneciente a colegios públicos, y otra perteneciente a colegios privados, si existen distintas tipologías de 'climas sociales en el aula', a partir de las percepciones de los alumnos, y si hay diferencias entre los centros públicos y los privados. 575 Sujetos, 200 alumnos de colegios públicos y 375 de colegios privados. Se trata de sujetos de ambos sexos, con edades entre 13 y 14 años, pertenecientes a un nivel de 8 de EGB de Avilés, Gijon y Oviedo. Variables independientes: implicación, afiliación, ayuda, tarea, competitividad, organización, claridad, control e innovación. Variables moduladoras: pertenencia por parte de los alumnos a colegios públicos o privados. Escala de clima social (ces), creada por r.H. Moos y cols.. Análisis de conglomerados cluster K-means, un tipo de análisis de cluster no jerárquico. Con este método se divide un conjunto de individuos en conglomerados, de tal forma que, al final del proceso, cada caso pertenece al cluster cuyo centro está más cercano a él. El centro del cluster viene dado por la media de los individuos que forman cada variable. Del análisis de variables que intervienen en la percepción del clima social escolar, se observan diferencias entre colegios públicos y privados, en lo que respecta a las variables de ayuda, tarea, organización e innovación. En relación a las otras cinco variables, afiliación, implicación, competitividad, claridad y control, las diferencias entre una muestra y otra son inexistentes. A la hora de estudiar cada uno de los cluster, se tiene en cuenta la reestructuración realizada tanto en la muestra de colegios públicos como privados. En la muestra de colegios privados destacan tres tipologías de climas: un clima afectivo percibido por un 50 por ciento de la población; un clima conservador y autoritario percibido por casi un 40 por ciento de los estudiantes; un clima estructurado percibido por un 10 por ciento aproximadamente. En la muestra de alumnos pertenecientes a colegios públicos, se encuentran cuatro tipos de climas: un clima afectivo percibido por un 32 por ciento de la población; un clima afectivo y no participativo, detectado por un 27 por ciento de los estudiantes; un clima autoritario percibido por un 26,5 por ciento de la muestra; un clima centrado en la organización y el esfuerzo, percibido por un 14,5 por ciento de la población. El hecho de que los estudiantes de colegios públicos o privados, perciban un determinado tipo de clima, está muy relacionado con la figura del profesor-tutor. El funcionamiento de la clase depende de las características de éste, que aunque revelen los canones de la institución, tienen una huella personal. Para evaluar la percepción del clima escolar, a las variables analizadas, habría que añadir la personalidad del profesor, lo que no descartan realizar en una posterior investigación.

Percepciones del clima social escolar : diferencias tipol??gicas entre varones y mujeres

Resumen tomado de la publicaci??n

Improving Clustering Algorithms for Image Segmentation using Contour and Region Information

In image segmentation, clustering algorithms are very popular because they are intuitive and, some of them, easy to implement. For instance, the k-means is one of the most used in the literature, and many authors successfully compare their new proposal with the results achieved by the k-means. However, it is well known that clustering image segmentation has many problems. For instance, the number of regions of the image has to be known a priori, as well as different initial seed placement (initial clusters) could produce different segmentation results. Most of these algorithms could be slightly improved by considering the coordinates of the image as features in the clustering process (to take spatial region information into account). In this paper we propose a significant improvement of clustering algorithms for image segmentation. The method is qualitatively and quantitative evaluated over a set of synthetic and real images, and compared with classical clustering approaches. Results demonstrate the validity of this new approach

A natureza na lineal dos funcionamentos inferenciais: um estudo empírico com base no humor gráfico

Analisa-se se os funcionamentos inferenciais têm uma estrutura própria dos sistemas dinâmicos não lineais, estudados a partir de quatro gráficas humorísticas. Os primeiros resultados com o tratamento estadístico lineal de K-medias projetam a presencia de perfis de diferentes funcionamentos inferenciais em função das diferentes piadas. Os resultados com a técnica da wavelet, proveniente dos sistemas dinâmicos não lineais, mostram patrões dos funcionamentos inferenciais que dão conta de sua natureza multifractal, sem uma sequencialidade fixa e sem uma organização aparente. Isto implica que é necessário revisar a concepção de estádios sequenciais fixos como os que dominam os estudos de desenvolvimento cognitivo.

Diferencias intergrupales en el sistema de metas personales, satisfacción y rendimiento en adolescentes.

Integrar diferentes unidades de análisis para el estudio de la personalidad y considerar estas unidades en su predicción de la satisfacción y el rendimiento en adolescentes. 296 estudiantes de ESO de entre 15 y 18 años. 162 son mujeres y 134 varones. Las aplicaciones de las pruebas se realizan en horario de tutorías dentro del Plan Acción Tutorial (PAT). Se les explica a los alumnos que participan en la investigación sobre 'metas que se proponen realizar en un futuro' y que las pruebas que se administran les pueden ayudar en el futuro para la toma de decisiones. Las aplicaciones de las pruebas se realizan en dos sesiones de evaluación. En la primera, se aplican las pruebas de personalidad y satisfacción. En la segunda se evalúan metas personales. El rendimiento académico se operativiza por la puntuación del adolescente en su curso académico. Todos los alumnos participan voluntariamente en la investigación. Escala de objetivos o metas personales, escala de satisfacción por áreas vitales (ESAV), Inventario de personalidad para adolescentes de Millón (MAPI), estilos básicos de personalidad, escalas de correlatos comportamentales. Para el análisis de los datos, se utilizan programas estadísticos SPSS, SPAD, LISREL VIII y para el cálculo del tamaño del efecto el Statistical Power Computer Analysis. Las técnicas de análisis de datos se centran en Análisis de Correspondencia Múltiple (ACM), análisis de conglomerados K means, Análisis de varianza y diferencias entre coeficientes de correlación. Los resultados indican que los adolescentes que se plantean metas relacionadas con las tareas vitales a desarrollar en un futuro próximo manifiestan mayores niveles de satisfacción. Además, las diferencias en los estilos de personalidad, permiten entender el sistema de metas personales en cuatro grupos de adolescentes. La consideración de los estilos de personalidad y las metas personales permiten entender la adaptación de los adolescentes a su entorno considerando la satisfacción autopercibida y el rendimiento académico.

La influencia del autoconcepto en el rendimiento académico en estudiantes universitarios. 'The influence of self-concept on academic achievement in university students'.

Resumen tomado de la publicación. Con el apoyo económico del departamento MIDE de la UNED. Incluye anexo con el cuestionario utilizado para la realización del estudio

Clasificaci??n de perfiles de uso de smartphones en estudiantes y docentes de la Universidad Aut??noma de Baja California, M??xico

Resumen tomado de la publicaci??n

A continuum of sudden stratospheric warmings

The k-means cluster technique is used to examine 43 yr of daily winter Northern Hemisphere (NH) polar stratospheric data from the 40-yr ECMWF Re-Analysis (ERA-40). The results show that the NH winter stratosphere exists in two natural well-separated states. In total, 10% of the analyzed days exhibit a warm disturbed state that is typical of sudden stratospheric warming events. The remaining 90% of the days are in a state typical of a colder undisturbed vortex. These states are determined objectively, with no preconceived notion of the groups. The two stratospheric states are described and compared with alternative indicators of the polar winter flow, such as the northern annular mode. It is shown that the zonally averaged zonal winds in the polar upper stratosphere at 7 hPa can best distinguish between the two states, using a threshold value of 4 m s−1, which is remarkably close to the standard WMO criterion for major warming events. The analysis also determines that there are no further divisions within the warm state, indicating that there is no well-designated threshold between major and minor warmings, nor between split and displaced vortex events. These different manifestations are simply members of a continuum of warming events.

Mean-tracking clustering algorithm for radial basis function centre selection

Radial basis functions can be combined into a network structure that has several advantages over conventional neural network solutions. However, to operate effectively the number and positions of the basis function centres must be carefully selected. Although no rigorous algorithm exists for this purpose, several heuristic methods have been suggested. In this paper a new method is proposed in which radial basis function centres are selected by the mean-tracking clustering algorithm. The mean-tracking algorithm is compared with k means clustering and it is shown that it achieves significantly better results in terms of radial basis function performance. As well as being computationally simpler, the mean-tracking algorithm in general selects better centre positions, thus providing the radial basis functions with better modelling accuracy

Givens rotation based fast backward elimination algorithm for RBF neural network pruning

A fast backward elimination algorithm is introduced based on a QR decomposition and Givens transformations to prune radial-basis-function networks. Nodes are sequentially removed using an increment of error variance criterion. The procedure is terminated by using a prediction risk criterion so as to obtain a model structure with good generalisation properties. The algorithm can be used to postprocess radial basis centres selected using a k-means routine and, in this mode, it provides a hybrid supervised centre selection approach.

Neural network basis function center selection using cluster analysis

This paper deals with the selection of centres for radial basis function (RBF) networks. A novel mean-tracking clustering algorithm is described as a way in which centers can be chosen based on a batch of collected data. A direct comparison is made between the mean-tracking algorithm and k-means clustering and it is shown how mean-tracking clustering is significantly better in terms of achieving an RBF network which performs accurate function modelling.

Non-uniform data distribution for communication-efficient parallel clustering

Global communicationrequirements andloadimbalanceof someparalleldataminingalgorithms arethe major obstacles to exploitthe computational power of large-scale systems. This work investigates how non-uniform data distributions can be exploited to remove the global communication requirement and to reduce the communication costin parallel data mining algorithms and, in particular, in the k-means algorithm for cluster analysis. In the straightforward parallel formulation of the k-means algorithm, data and computation loads are uniformly distributed over the processing nodes. This approach has excellent load balancing characteristics that may suggest it could scale up to large and extreme-scale parallel computing systems. However, at each iteration step the algorithm requires a global reduction operationwhichhinders thescalabilityoftheapproach.Thisworkstudiesadifferentparallelformulation of the algorithm where the requirement of global communication is removed, while maintaining the same deterministic nature ofthe centralised algorithm. The proposed approach exploits a non-uniform data distribution which can be either found in real-world distributed applications or can be induced by means ofmulti-dimensional binary searchtrees. The approachcanalso be extended to accommodate an approximation error which allows a further reduction ofthe communication costs. The effectiveness of the exact and approximate methods has been tested in a parallel computing system with 64 processors and in simulations with 1024 processing element

Estimation of wind storm impacts over Western Germany under future climate conditions using a statistical-dynamical downscaling approach

A statistical–dynamical regionalization approach is developed to assess possible changes in wind storm impacts. The method is applied to North Rhine-Westphalia (Western Germany) using the FOOT3DK mesoscale model for dynamical downscaling and ECHAM5/OM1 global circulation model climate projections. The method first classifies typical weather developments within the reanalysis period using K-means cluster algorithm. Most historical wind storms are associated with four weather developments (primary storm-clusters). Mesoscale simulations are performed for representative elements for all clusters to derive regional wind climatology. Additionally, 28 historical storms affecting Western Germany are simulated. Empirical functions are estimated to relate wind gust fields and insured losses. Transient ECHAM5/OM1 simulations show an enhanced frequency of primary storm-clusters and storms for 2060–2100 compared to 1960–2000. Accordingly, wind gusts increase over Western Germany, reaching locally +5% for 98th wind gust percentiles (A2-scenario). Consequently, storm losses are expected to increase substantially (+8% for A1B-scenario, +19% for A2-scenario). Regional patterns show larger changes over north-eastern parts of North Rhine-Westphalia than for western parts. For storms with return periods above 20 yr, loss expectations for Germany may increase by a factor of 2. These results document the method's functionality to assess future changes in loss potentials in regional terms.

Extreme wind storms over Europe in present and future climate: a cluster analysis approach

Boreal winter wind storm situations over Central Europe are investigated by means of an objective cluster analysis. Surface data from the NCEP-Reanalysis and ECHAM4/OPYC3-climate change GHG simulation (IS92a) are considered. To achieve an optimum separation of clusters of extreme storm conditions, 55 clusters of weather patterns are differentiated. To reduce the computational effort, a PCA is initially performed, leading to a data reduction of about 98 %. The clustering itself was computed on 3-day periods constructed with the first six PCs using "k-means" clustering algorithm. The applied method enables an evaluation of the time evolution of the synoptic developments. The climate change signal is constructed by a projection of the GCM simulation on the EOFs attained from the NCEP-Reanalysis. Consequently, the same clusters are obtained and frequency distributions can be compared. For Central Europe, four primary storm clusters are identified. These clusters feature almost 72 % of the historical extreme storms events and add only to 5 % of the total relative frequency. Moreover, they show a statistically significant signature in the associated wind fields over Europe. An increased frequency of Central European storm clusters is detected with enhanced GHG conditions, associated with an enhancement of the pressure gradient over Central Europe. Consequently, more intense wind events over Central Europe are expected. The presented algorithm will be highly valuable for the analysis of huge data amounts as is required for e.g. multi-model ensemble analysis, particularly because of the enormous data reduction.