Generation of human computational models with machine learning

Services in smart environments pursue to increase the quality of people?s lives. The most important issues when developing this kind of environments is testing and validating such services. These tasks usually imply high costs and annoying or unfeasible real-world testing. In such cases, artificial societies may be used to simulate the smart environment (i.e. physical environment, equipment and humans). With this aim, the CHROMUBE methodology guides test engineers when modeling human beings. Such models reproduce behaviors which are highly similar to the real ones. Originally, these models are based on automata whose transitions are governed by random variables. Automaton?s structure and the probability distribution functions of each random variable are determined by a manual test and error process. In this paper, it is presented an alternative extension of this methodology which avoids the said manual process. It is based on learning human behavior patterns automatically from sensor data by using machine learning techniques. The presented approach has been tested on a real scenario, where this extension has given highly accurate human behavior models,

Evolution over time of heavy vehicle volume in toll roads: A dynamic panel data to identify key explanatory variables in Spain

Improving the knowledge of demand evolution over time is a key aspect in the evaluation of transport policies and in forecasting future investment needs. It becomes even more critical for the case of toll roads, which in recent decades has become an increasingly common device to fund road projects. However, literature regarding demand elasticity estimates in toll roads is sparse and leaves some important aspects to be analyzed in greater detail. In particular, previous research on traffic analysis does not often disaggregate heavy vehicle demand from the total volume, so that the specific behavioral patternsof this traffic segment are not taken into account. Furthermore, GDP is the main socioeconomic variable most commonly chosen to explain road freight traffic growth over time. This paper seeks to determine the variables that better explain the evolution of heavy vehicle demand in toll roads over time. To that end, we present a dynamic panel data methodology aimed at identifying the key socioeconomic variables that explain the behavior of road freight traffic throughout the years. The results show that, despite the usual practice, GDP may not constitute a suitable explanatory variable for heavy vehicle demand. Rather, considering only the GDP of those sectors with a high impact on transport demand, such as construction or industry, leads to more consistent results. The methodology is applied to Spanish toll roads for the 1990?2011 period. This is an interesting case in the international context, as road freight demand has experienced an even greater reduction in Spain than elsewhere, since the beginning of the economic crisis in 2008.

Explaining light vehicle demand evolution in interurban toll roads: a dynamic panel data analysis in Spain

Tolls have increasingly become a common mechanism to fund road projects in recent decades. Therefore, improving knowledge of demand behavior constitutes a key aspect for stakeholders dealing with the management of toll roads. However, the literature concerning demand elasticity estimates for interurban toll roads is still limited due to their relatively scarce number in the international context. Furthermore, existing research has left some aspects to be investigated, among others, the choice of GDP as the most common socioeconomic variable to explain traffic growth over time. This paper intends to determine the variables that better explain the evolution of light vehicle demand in toll roads throughout the years. To that end, we establish a dynamic panel data methodology aimed at identifying the key socioeconomic variables explaining changes in light vehicle demand over time. The results show that, despite some usefulness, GDP does not constitute the most appropriate explanatory variable, while other parameters such as employment or GDP per capita lead to more stable and consistent results. The methodology is applied to Spanish toll roads for the 1990?2011 period, which constitutes a very interesting case on variations in toll road use, as road demand has experienced a significant decrease since the beginning of the economic crisis in 2008.

Flexibility in Airline Business Models with Core Competence as an Indicator.

The airline industry is often unstable and unpredictable forcing airlines to restructure and create flexible strategies that can respond to external operating environmental changes. In turbulent and competitive environments, firms with higher flexibility perform better and the value of these flexibilities depends on factors of uncertainty in the competitive environment. A model is sought for and arrived at, that shows how an airline business model will function in an uncertain environment with the least reduction in business performance over time. An analysis of the business model flexibility of 17 Airlines from Asia, Europe and Oceania, that is done with core competence as the indicator reveals a picture of inconsistencies in the core competence strategy of certain airlines and the corresponding reduction in business performance. The performance variations are explained from a service oriented core competence strategy employed by airlines that ultimately enables them in having a flexible business model that not only increases business performance but also helps in reducing the uncertainties in the internal and external operating environments.

Atmospheric attenuation at 100 and 300 GHz estimated with radiosonde data

The study of atmospheric propagation impairments at submillimeter and THz frequencies is becoming increasingly relevant due to the strong effects caused by the composition of the troposphere and the phenomena occurring in it. The present paper is devoted to the estimation of total attenuation at 100 GHz and 300 GHz under non-rainy scenarios. With this purpose, 4 years of meteorological data from Madrid have been collected, including radiosoundings from Madrid-Barajas Airport and co-site SYNOP observations. This volume of data has been analyzed with the aim of also introducing a detection method of rain conditions, which cannot be easily identified in radiosounding profiles. Finally, the method has been used to discard several probable events which would be responsible of scattering conditions and, hence, yearly CDFs of total attenuation have been obtained. It is expected that the statistics would be closest to the ones obtained by experimental techniques under similar atmospheric conditions.

Hybrid analysis of nonlinear circuits: DAE models with indices zero and one

We extend in this paper some previous results concerning the differential-algebraic index of hybrid models of electrical and electronic circuits. Specifically, we present a comprehensive index characterization which holds without passivity requirements, in contrast to previous approaches, and which applies to nonlinear circuits composed of uncoupled, one-port devices. The index conditions, which are stated in terms of the forest structure of certain digraph minors, do not depend on the specific tree chosen in the formulation of the hybrid equations. Additionally, we show how to include memristors in hybrid circuit models; in this direction, we extend the index analysis to circuits including active memristors, which have been recently used in the design of nonlinear oscillators and chaotic circuits. We also discuss the extension of these results to circuits with controlled sources, making our framework of interest in the analysis of circuits with transistors, amplifiers, and other multiterminal devices.

Implications of agricultural production, policy and land use changes on water resource assessment

Una gestión más eficiente y equitativa del agua a escala de cuenca no se puede centrar exclusivamente en el recurso hídrico en sí, sino también en otras políticas y disciplinas científicas. Existe un consenso creciente de que, además de la consideración de las cambiantes condiciones climáticas, es necesaria una integración de ámbitos de investigación tales como la agronomía, planificación del territorio y ciencias políticas y económicas a fin de satisfacer de manera sostenible las demandas de agua por parte de la sociedad y del medio natural. La Política Agrícola Común (PAC) es el principal motor de cambio en las tendencias de paisajes rurales y sistemas agrícolas, pero el deterioro del medio ambiente es ahora una de las principales preocupaciones. Uno de los cambios más relevantes se ha producido con la expansión e intensificación del olivar en España, principalmente con nuevas zonas de regadío o la conversión de olivares de secano a sistemas en regadío. Por otra parte, el cambio de las condiciones climáticas podría ejercer un papel importante en las tendencias negativas de las aportaciones a los ríos, pero no queda claro el papel que podrían estar jugando los cambios de uso de suelo y cobertura vegetal sobre las tendencias negativas de caudal observadas. Esta tesis tiene como objetivo mejorar el conocimiento de los efectos de la producción agrícola, política agraria y cambios de uso de suelo y cobertura vegetal sobre las condiciones de calidad del agua, respuesta hidrológica y apropiación del agua por parte de la sociedad. En primer lugar, el estudio determina las tendencias existentes de nitratos y sólidos en suspensión en las aguas superficiales de la cuenca del río Guadalquivir durante el periodo de 1998 a 2009. Desde una perspectiva de política agraria, la investigación trata de evaluar mediante un análisis de datos de panel las principales variables, incluyendo la reforma de la PAC de 2003, que están teniendo una influencia en ambos indicadores de calidad. En segundo lugar, la apropiación del agua y el nivel de contaminación por nitratos debido a la producción del aceite de oliva en España se determinan con una evaluación de la huella hídrica (HH), teniendo en cuenta una variabilidad espacial y temporal a largo de las provincias españolas y entre 1997 y 2008. Por último, la tesis analiza los efectos de los cambios de uso de suelo y cobertura vegetal sobre las tendencias negativas observadas en la zona alta del Turia, cabecera de la cuenca del río Júcar, durante el periodo 1973-2008 mediante una modelización ecohidrológica. En la cuenca del Guadalquivir cerca del 20% de las estaciones de monitoreo muestran tendencias significativas, lineales o cuadráticas, para cada indicador de calidad de agua. La mayoría de las tendencias significativas en nitratos están aumentando, y la mayoría de tendencias cuadráticas muestran un patrón en forma de U. Los modelos de regresión de datos de panel muestran que las variables más importantes que empeoran ambos indicadores de calidad del agua son la intensificación de biomasa y las exportaciones de ambos indicadores de calidad procedentes de aguas arriba. En regiones en las que el abandono agrícola y/o desintensificación han tenido lugar han mejorado las condiciones de calidad del agua. Para los nitratos, el desacoplamiento de las subvenciones a la agricultura y la reducción de la cuantía de las subvenciones a tierras de regadío subyacen en la reducción observada de la concentración de nitratos. Las medidas de modernización de regadíos y el establecimiento de zonas vulnerables a nitratos reducen la concentración en subcuencas que muestran una tendencia creciente de nitratos. Sin embargo, el efecto de las exportaciones de nitratos procedente de aguas arriba, la intensificación de la biomasa y los precios de los cultivos presentan un mayor peso, explicando la tendencia creciente observada de nitratos. Para los sólidos en suspensión, no queda de forma evidente si el proceso de desacoplamiento ha influido negativa o positivamente. Sin embargo, los mayores valores de las ayudas agrarias aún ligadas a la producción, en particular en zonas de regadío, conllevan un aumento de las tasas de erosión. Aunque la cuenca del Guadalquivir ha aumentado la producción agrícola y la eficiencia del uso del agua, el problema de las altas tasas de erosión aún no ha sido mitigado adecuadamente. El estudio de la huella hídrica (HH) revela que en 1 L de aceite de oliva español más del 99,5% de la HH está relacionado con la producción de la aceituna, mientras que menos del 0,5% se debe a otros componentes, es decir, a la botella, tapón y etiqueta. Durante el período estudiado, la HH verde en secano y en regadío representa alrededor del 72% y 12%, respectivamente, del total de la HH. Las HHs azul y gris representan 6% y 10%, respectivamente. La producción de aceitunas se concentra en regiones con una HH menor por unidad de producto. La producción de aceite de oliva ha aumentado su productividad del agua durante 1997-2008, incentivado por los crecientes precios del aceite, como también lo ha hecho la cantidad de exportaciones de agua virtual. De hecho, las mayores zonas productoras presentan una eficiencia alta del uso y de productividad del agua, así como un menor potencial de contaminación por nitratos. Pero en estas zonas se ve a la vez reflejado un aumento de presión sobre los recursos hídricos locales. El aumento de extracciones de agua subterránea relacionadas con las exportaciones de aceite de oliva podría añadir una mayor presión a la ya estresada cuenca del Guadalquivir, mostrando la necesidad de equilibrar las fuerzas del mercado con los recursos locales disponibles. Los cambios de uso de suelo y cobertura vegetal juegan un papel importante en el balance del agua de la cuenca alta del Turia, pero no son el principal motor que sustenta la reducción observada de caudal. El aumento de la temperatura es el principal factor que explica las mayores tasas de evapotranspiración y la reducción de caudales. Sin embargo, los cambios de uso de suelo y el cambio climático han tenido un efecto compensatorio en la respuesta hidrológica. Por un lado, el caudal se ha visto afectado negativamente por el aumento de la temperatura, mientras que los cambios de uso de suelo y cobertura vegetal han compensado positivamente con una reducción de las tasas de evapotranspiración, gracias a los procesos de disminución de la densidad de matorral y de degradación forestal. El estudio proporciona una visión que fortalece la interdisciplinariedad entre la planificación hidrológica y territorial, destacando la necesidad de incluir las implicaciones de los cambios de uso de suelo y cobertura vegetal en futuros planes hidrológicos. Estos hallazgos son valiosos para la gestión de la cuenca del río Turia, y el enfoque empleado es útil para la determinación del peso de los cambios de uso de suelo y cobertura vegetal en la respuesta hidrológica en otras regiones. ABSTRACT Achieving a more efficient and equitable water management at catchment scale does not only rely on the water resource itself, but also on other policies and scientific knowledge. There is a growing consensus that, in addition to consideration of changing climate conditions, integration with research areas such as agronomy, land use planning and economics and political science is required to meet sustainably the societal and environmental water demands. The Common Agricultural Policy (CAP) is a main driver for trends in rural landscapes and agricultural systems, but environmental deterioration is now a principal concern. One of the most relevant changes has occurred with the expansion and intensification of olive orchards in Spain, taking place mainly with new irrigated areas or with the conversion from rainfed to irrigated systems. Moreover, changing climate conditions might exert a major role on water yield trends, but it remains unclear the role that ongoing land use and land cover changes (LULCC) might have on observed river flow trends. This thesis aims to improve the understanding of the effects of agricultural production, policies and LULCC on water quality conditions, hydrological response and human water appropriation. Firstly, the study determines the existing trends for nitrates and suspended solids in the Guadalquivir river basin’s surface waters (south Spain) during the period from 1998 to 2009. From a policy perspective, the research tries to assess with panel data analysis the main drivers, including the 2003 CAP reform, which are having an influence on both water quality indicators. Secondly, water appropriation and nitrate pollution level originating from the production of olive oil in Spain is determined with a water footprint (WF) assessment, considering a spatial temporal variability across the Spanish provinces and from 1997 to 2008 years. Finally, the thesis analyzes the effects of the LULCC on the observed negative trends over the period 1973-2008 in the Upper Turia basin, headwaters of the Júcar river demarcation (east Spain), with ecohydrological modeling. In the Guadalquivir river basin about 20% of monitoring stations show significant trends, linear or quadratic, for each water quality indicator. Most significant trends of nitrates are augmenting than decreasing, and most significant quadratic terms of both indicators exhibit U-shaped patterns. The panel data models show that the most important drivers that are worsening nitrates and suspended solids in the basin are biomass intensification and exports of both water quality indicators from upland regions. In regions that agricultural abandonment and/or de-intensification have taken place the water quality conditions have improved. For nitrates, the decoupling of agricultural subsidies and the reduction of the amount of subsidies to irrigated land underlie the observed reduction of nitrates concentration. Measures of irrigation modernization and establishment of vulnerable zones to nitrates ameliorate the concentration of nitrates in subbasins showing an increasing trend. However, the effect of nitrates load from upland areas, intensification of biomass and crop prices present a greater weight leading to the final increasing trend in this subbasins group, where annual crops dominate. For suspended solids, there is no clear evidence that decoupling process have influenced negatively or positively. Nevertheless, greater values of subsidies still linked to production, particularly in irrigated regions, lead to increasing erosion rates. Although agricultural production has augmented in the basin and water efficiency in the agricultural sector has improved, the issue of high erosion rates has not yet been properly faced. The water footprint (WF) assessment reveals that for 1 L Spanish olive oil more than 99.5% of the WF is related to the olive fruit production, whereas less than 0.5% is due to other components i.e. bottle, cap and label. Over the studied period, the green WF in rainfed and irrigated systems represents about 72% and 12%, respectively, of the total WF. Blue and grey WFs represent 6% and 10%, respectively. The olive production is concentrated in regions with the smallest WF per unit of product. The olive oil production has increased its apparent water productivity from 1997 to 2008 incentivized by growing trade prices, but also did the amount of virtual water exports. In fact, the largest producing areas present high water use efficiency per product and apparent water productivity as well as less nitrates pollution potential, but this enhances the pressure on the available water resources. Increasing groundwater abstractions related to olive oil exports may add further pressure to the already stressed Guadalquivir basin. This shows the need to balance the market forces with the available local resources. Concerning the effects of LULCC on the Upper Turia basin’s streamflow, LULCC play a significant role on the water balance, but it is not the main driver underpinning the observed reduction on Turia's streamflow. Increasing mean temperature is the main factor supporting larger evapotranspiration rates and streamflow reduction. In fact, LULCC and climate change have had an offsetting effect on the streamflow generation during the study period. While streamflow has been negatively affected by increasing temperature, ongoing LULCC have positively compensated with reduced evapotranspiration rates, thanks to mainly shrubland clearing and forest degradation processes. The research provides insight for strengthening the interdisciplinarity between hydrological and spatial planning, highlighting the need to include the implications of LULCC in future hydrological plans. These findings are valuable for the management of the Turia river basin, as well as a useful approach for the determination of the weight of LULCC on the hydrological response in other regions.

Regression models with MoPs Bayesian networks

We present a model of Bayesian network for continuous variables, where densities and conditional densities are estimated with B-spline MoPs. We use a novel approach to directly obtain conditional densities estimation using B-spline properties. In particular we implement naive Bayes and wrapper variables selection. Finally we apply our techniques to the problem of predicting neurons morphological variables from electrophysiological ones.

Analysis of the impact of globalization and economic growth on food security in developing countries

A pesar de los importantes avances en la reducción del hambre, la seguridad alimentaria continúa siendo un reto de dimensión internacional. La seguridad alimentaria es un concepto amplio y multidimensional, cuyo análisis abarca distintas escalas y horizontes temporales. Dada su complejidad, la identificación de las causas de la inseguridad alimentaria y la priorización de las medias para abordarlas, son dos cuestiones que suscitan un intenso debate en la actualidad. El objetivo de esta tesis es evaluar el impacto de la globalización y el crecimiento económico en la seguridad alimentaria en los países en desarrollo, desde una perspectiva macro y un horizonte temporal a largo plazo. La influencia de la globalización se aborda de una manera secuencial. En primer lugar, se analiza la relación entre la inversión público-privada en infraestructuras y las exportaciones agrarias. A continuación, se estudia el impacto de las exportaciones agrarias en los indicadores de seguridad alimentaria. El estudio del impacto del crecimiento económico aborda los cambios paralelos en la distribución de la renta, y cómo la inequidad influye en el comportamiento de la seguridad alimentaria nacional. Además, se analiza en qué medida el crecimiento económico contribuye a acelerar el proceso de mejora de la seguridad alimentaria. Con el fin de conseguir los objetivos mencionados, se llevan a cabo varios análisis econométricos basados en datos de panel, en el que se combinan datos de corte transversal de 52 países y datos temporales comprendidos en el periodo 1991-2012. Se analizan tanto variables en niveles como variables en tasas de cambio anual. Se aplican los modelos de estimación de efectos variables y efectos fijos, ambos en niveles y en primeras diferencias. La tesis incluye cuatro tipos de modelos econométricos, cada uno de ellos con sus correspondientes pruebas de robustez y especificaciones. Los resultados matizan la importancia de la globalización y el crecimiento económico como mecanismos de mejora de la seguridad alimentaria en los países en desarrollo. Se obtienen dos conclusiones relativas a la globalización. En primer lugar, los resultados sugieren que la promoción de las inversiones privadas en infraestructuras contribuye a aumentar las exportaciones agrarias. En segundo lugar, se observa que las exportaciones agrarias pueden tener un impacto negativo en los indicadores de seguridad alimentaria. La combinación de estas dos conclusiones sugiere que la apertura comercial y financiera no contribuye por sí misma a la mejora de la seguridad alimentaria en los países en desarrollo. La apertura internacional de los países en desarrollo ha de ir acompañada de políticas e inversiones que desarrollen sectores productivos de alto valor añadido, que fortalezcan la economía nacional y reduzcan su dependencia exterior. En relación al crecimiento económico, a pesar del incuestionable hecho de que el crecimiento económico es una condición necesaria para reducir los niveles de subnutrición, no es una condición suficiente. Se han identificado tres estrategias adicionales que han de acompañar al crecimiento económico con el fin de intensificar su impacto positivo sobre la subnutrición. Primero, es necesario que el crecimiento económico sea acompañado de una distribución más equitativa de los ingresos. Segundo, el crecimiento económico ha de reflejarse en un aumento de inversiones en salud, agua y saneamiento y educación. Se observa que, incluso en ausencia de crecimiento económico, mejoras en el acceso a agua potable contribuyen a reducir los niveles de población subnutrida. Tercero, el crecimiento económico sostenible en el largo plazo parece tener un mayor impacto positivo sobre la seguridad alimentaria que el crecimiento económico más volátil o inestable en el corto plazo. La estabilidad macroeconómica se identifica como una condición necesaria para alcanzar una mayor mejora en la seguridad alimentaria, incluso habiéndose mejorado la equidad en la distribución de los ingresos. Por último, la tesis encuentra que los países en desarrollo analizados han experimentado diferentes trayectorias no lineales en su proceso de mejora de sus niveles de subnutrición. Los resultados sugieren que un mayor nivel inicial de subnutrición y el crecimiento económico son responsables de una respuesta más rápida al reto de la mejora de la seguridad alimentaria. ABSTRACT Despite the significant reductions of hunger, food security still remains a global challenge. Food security is a wide concept that embraces multiple dimensions, and has spatial-temporal scales. Because of its complexity, the identification of the drivers underpinning food insecurity and the prioritization of measures to address them are a subject of intensive debate. This thesis attempts to assess the impact of globalization and economic growth on food security in developing countries with a macro level scale (country) and using a long-term approach. The influence of globalization is addressed in a sequential way. First, the impact of public-private investment in infrastructure on agricultural exports in developing countries is analyzed. Secondly, an assessment is conducted to determine the impact of agricultural exports on food security indicators. The impact of economic growth focuses on the parallel changes in income inequality and how the income distribution influences countries' food security performance. Furthermore, the thesis analyzes to what extent economic growth helps accelerating food security improvements. To address the above mentioned goals, various econometric models are formulated. Models use panel data procedures combining cross-sectional data of 52 countries and time series data from 1991 to 2012. Yearly data are expressed both in levels and in changes. The estimation models applied are random effects estimation and fixed effects estimations, both in levels and in first differences. The thesis includes four families of econometric models, each with its own set of robustness checks and specifications. The results qualify the relevance of globalization and economic growth as enabling mechanisms for improving food security in developing countries. Concerning globalization, two main conclusions can be drawn. First, results showed that enhancing foreign private investment in infrastructures contributes to increase agricultural exports. Second, agricultural exports appear to have a negative impact on national food security indicators. These two conclusions suggest that trade and financial openness per se do not contribute directly to improve food security in development countries. Both measures should be accompanied by investments and policies to support the development of national high value productive sectors, to strengthen the domestic economy and reduce its external dependency. Referring to economic growth, despite the unquestionable fact that income growth is a pre-requisite for reducing undernourishment, results suggest that it is a necessary but not a sufficient condition. Three additional strategies should accompany economic growth to intensifying its impact on food security. Firstly, it is necessary that income growth should be accompanied by a better distribution of income. Secondly, income growth needs to be followed by investments and policies in health, sanitation and education to improve food security. Even if economic growth falters, sustained improvements in the access to drinking water may still give rise to reductions in the percentage of undernourished people. And thirdly, long-term economic growth appears to have a greater impact on reducing hunger than growth regimes that combine periods of growth peaks followed by troughs. Macroeconomic stability is a necessary condition for accelerating food security. Finally, the thesis finds that the developing countries analyzed have experienced different non-linear paths toward improving food security. Results also show that a higher initial level of undernourishment and economic growth result in a faster response for improving food security.

A comparison of open-source segmentation architectures for dealing with imperfect data from the media in speech synthesis

Traditional Text-To-Speech (TTS) systems have been developed using especially-designed non-expressive scripted recordings. In order to develop a new generation of expressive TTS systems in the Simple4All project, real recordings from the media should be used for training new voices with a whole new range of speaking styles. However, for processing this more spontaneous material, the new systems must be able to deal with imperfect data (multi-speaker recordings, background and foreground music and noise), filtering out low-quality audio segments and creating mono-speaker clusters. In this paper we compare several architectures for combining speaker diarization and music and noise detection which improve the precision and overall quality of the segmentation.

Estudio de la fisuración de paneles sándwich de yeso laminado y alma de lana de roca = Study of the cracking of sandwich panels of plasterboard and rockwool

Este artículo presenta el estudio de la rotura de paneles sándwich de yeso laminado y lana de roca bajo solicitaciones de flexo-tracción dentro de su plano. Estos paneles se emplean para conformar tabiques interiores de edificación y con frecuencia se fisuran por flechas excesivas en los forjados. Actualmente no hay modelos de cálculo fiables ni datos experimentales que permitan estudiar este problema. Este trabajo presenta los resultados de una campaña experimental encaminada a caracterizar el comportamiento en rotura de los paneles sándwich y de sus componentes individuales. Además, se presenta un modelo cohesivo con fisura embebida que permite simular el comportamiento en rotura del panel sándwich conjunto. Por último se presentan los resultados de los ensayos de fractura en modo mixto (tracción/cortante) de paneles comerciales y se reproduce su comportamiento con el modelo cohesivo propuesto, obteniéndose un buen ajuste. This paper presents the study of plasterboard and rockwool sandwich panels cracking under flexural loading. These panels are usually used to perform interior partition walls and they frequently show cracking pathology due to excessive deflexion of the slabs. There are currently no reliable simulation models and experimental data for the study of this problem. This paper presents the results of an experimental campaign aimed to characterize the fracture behaviour of sandwich panels and their individual components. In addition, the paper presents a cohesive model with embedded crack to simulate the fracture behaviour of the panel. Finally we present the results of tests for mixed mode fracture (tensile / shear) commercial panels and their behaviour is reproduced with the cohesive model proposed, yielding a good fit.

Multi-dimensional classification using Bayesian networks for stationary and evolving streaming data

Hoy en día, con la evolución continua y rápida de las tecnologías de la información y los dispositivos de computación, se recogen y almacenan continuamente grandes volúmenes de datos en distintos dominios y a través de diversas aplicaciones del mundo real. La extracción de conocimiento útil de una cantidad tan enorme de datos no se puede realizar habitualmente de forma manual, y requiere el uso de técnicas adecuadas de aprendizaje automático y de minería de datos. La clasificación es una de las técnicas más importantes que ha sido aplicada con éxito a varias áreas. En general, la clasificación se compone de dos pasos principales: en primer lugar, aprender un modelo de clasificación o clasificador a partir de un conjunto de datos de entrenamiento, y en segundo lugar, clasificar las nuevas instancias de datos utilizando el clasificador aprendido. La clasificación es supervisada cuando todas las etiquetas están presentes en los datos de entrenamiento (es decir, datos completamente etiquetados), semi-supervisada cuando sólo algunas etiquetas son conocidas (es decir, datos parcialmente etiquetados), y no supervisada cuando todas las etiquetas están ausentes en los datos de entrenamiento (es decir, datos no etiquetados). Además, aparte de esta taxonomía, el problema de clasificación se puede categorizar en unidimensional o multidimensional en función del número de variables clase, una o más, respectivamente; o también puede ser categorizado en estacionario o cambiante con el tiempo en función de las características de los datos y de la tasa de cambio subyacente. A lo largo de esta tesis, tratamos el problema de clasificación desde tres perspectivas diferentes, a saber, clasificación supervisada multidimensional estacionaria, clasificación semisupervisada unidimensional cambiante con el tiempo, y clasificación supervisada multidimensional cambiante con el tiempo. Para llevar a cabo esta tarea, hemos usado básicamente los clasificadores Bayesianos como modelos. La primera contribución, dirigiéndose al problema de clasificación supervisada multidimensional estacionaria, se compone de dos nuevos métodos de aprendizaje de clasificadores Bayesianos multidimensionales a partir de datos estacionarios. Los métodos se proponen desde dos puntos de vista diferentes. El primer método, denominado CB-MBC, se basa en una estrategia de envoltura de selección de variables que es voraz y hacia delante, mientras que el segundo, denominado MB-MBC, es una estrategia de filtrado de variables con una aproximación basada en restricciones y en el manto de Markov. Ambos métodos han sido aplicados a dos problemas reales importantes, a saber, la predicción de los inhibidores de la transcriptasa inversa y de la proteasa para el problema de infección por el virus de la inmunodeficiencia humana tipo 1 (HIV-1), y la predicción del European Quality of Life-5 Dimensions (EQ-5D) a partir de los cuestionarios de la enfermedad de Parkinson con 39 ítems (PDQ-39). El estudio experimental incluye comparaciones de CB-MBC y MB-MBC con los métodos del estado del arte de la clasificación multidimensional, así como con métodos comúnmente utilizados para resolver el problema de predicción de la enfermedad de Parkinson, a saber, la regresión logística multinomial, mínimos cuadrados ordinarios, y mínimas desviaciones absolutas censuradas. En ambas aplicaciones, los resultados han sido prometedores con respecto a la precisión de la clasificación, así como en relación al análisis de las estructuras gráficas que identifican interacciones conocidas y novedosas entre las variables. La segunda contribución, referida al problema de clasificación semi-supervisada unidimensional cambiante con el tiempo, consiste en un método nuevo (CPL-DS) para clasificar flujos de datos parcialmente etiquetados. Los flujos de datos difieren de los conjuntos de datos estacionarios en su proceso de generación muy rápido y en su aspecto de cambio de concepto. Es decir, los conceptos aprendidos y/o la distribución subyacente están probablemente cambiando y evolucionando en el tiempo, lo que hace que el modelo de clasificación actual sea obsoleto y deba ser actualizado. CPL-DS utiliza la divergencia de Kullback-Leibler y el método de bootstrapping para cuantificar y detectar tres tipos posibles de cambio: en las predictoras, en la a posteriori de la clase o en ambas. Después, si se detecta cualquier cambio, un nuevo modelo de clasificación se aprende usando el algoritmo EM; si no, el modelo de clasificación actual se mantiene sin modificaciones. CPL-DS es general, ya que puede ser aplicado a varios modelos de clasificación. Usando dos modelos diferentes, el clasificador naive Bayes y la regresión logística, CPL-DS se ha probado con flujos de datos sintéticos y también se ha aplicado al problema real de la detección de código malware, en el cual los nuevos ficheros recibidos deben ser continuamente clasificados en malware o goodware. Los resultados experimentales muestran que nuestro método es efectivo para la detección de diferentes tipos de cambio a partir de los flujos de datos parcialmente etiquetados y también tiene una buena precisión de la clasificación. Finalmente, la tercera contribución, sobre el problema de clasificación supervisada multidimensional cambiante con el tiempo, consiste en dos métodos adaptativos, a saber, Locally Adpative-MB-MBC (LA-MB-MBC) y Globally Adpative-MB-MBC (GA-MB-MBC). Ambos métodos monitorizan el cambio de concepto a lo largo del tiempo utilizando la log-verosimilitud media como métrica y el test de Page-Hinkley. Luego, si se detecta un cambio de concepto, LA-MB-MBC adapta el actual clasificador Bayesiano multidimensional localmente alrededor de cada nodo cambiado, mientras que GA-MB-MBC aprende un nuevo clasificador Bayesiano multidimensional. El estudio experimental realizado usando flujos de datos sintéticos multidimensionales indica los méritos de los métodos adaptativos propuestos. ABSTRACT Nowadays, with the ongoing and rapid evolution of information technology and computing devices, large volumes of data are continuously collected and stored in different domains and through various real-world applications. Extracting useful knowledge from such a huge amount of data usually cannot be performed manually, and requires the use of adequate machine learning and data mining techniques. Classification is one of the most important techniques that has been successfully applied to several areas. Roughly speaking, classification consists of two main steps: first, learn a classification model or classifier from an available training data, and secondly, classify the new incoming unseen data instances using the learned classifier. Classification is supervised when the whole class values are present in the training data (i.e., fully labeled data), semi-supervised when only some class values are known (i.e., partially labeled data), and unsupervised when the whole class values are missing in the training data (i.e., unlabeled data). In addition, besides this taxonomy, the classification problem can be categorized into uni-dimensional or multi-dimensional depending on the number of class variables, one or more, respectively; or can be also categorized into stationary or streaming depending on the characteristics of the data and the rate of change underlying it. Through this thesis, we deal with the classification problem under three different settings, namely, supervised multi-dimensional stationary classification, semi-supervised unidimensional streaming classification, and supervised multi-dimensional streaming classification. To accomplish this task, we basically used Bayesian network classifiers as models. The first contribution, addressing the supervised multi-dimensional stationary classification problem, consists of two new methods for learning multi-dimensional Bayesian network classifiers from stationary data. They are proposed from two different points of view. The first method, named CB-MBC, is based on a wrapper greedy forward selection approach, while the second one, named MB-MBC, is a filter constraint-based approach based on Markov blankets. Both methods are applied to two important real-world problems, namely, the prediction of the human immunodeficiency virus type 1 (HIV-1) reverse transcriptase and protease inhibitors, and the prediction of the European Quality of Life-5 Dimensions (EQ-5D) from 39-item Parkinson’s Disease Questionnaire (PDQ-39). The experimental study includes comparisons of CB-MBC and MB-MBC against state-of-the-art multi-dimensional classification methods, as well as against commonly used methods for solving the Parkinson’s disease prediction problem, namely, multinomial logistic regression, ordinary least squares, and censored least absolute deviations. For both considered case studies, results are promising in terms of classification accuracy as well as regarding the analysis of the learned MBC graphical structures identifying known and novel interactions among variables. The second contribution, addressing the semi-supervised uni-dimensional streaming classification problem, consists of a novel method (CPL-DS) for classifying partially labeled data streams. Data streams differ from the stationary data sets by their highly rapid generation process and their concept-drifting aspect. That is, the learned concepts and/or the underlying distribution are likely changing and evolving over time, which makes the current classification model out-of-date requiring to be updated. CPL-DS uses the Kullback-Leibler divergence and bootstrapping method to quantify and detect three possible kinds of drift: feature, conditional or dual. Then, if any occurs, a new classification model is learned using the expectation-maximization algorithm; otherwise, the current classification model is kept unchanged. CPL-DS is general as it can be applied to several classification models. Using two different models, namely, naive Bayes classifier and logistic regression, CPL-DS is tested with synthetic data streams and applied to the real-world problem of malware detection, where the new received files should be continuously classified into malware or goodware. Experimental results show that our approach is effective for detecting different kinds of drift from partially labeled data streams, as well as having a good classification performance. Finally, the third contribution, addressing the supervised multi-dimensional streaming classification problem, consists of two adaptive methods, namely, Locally Adaptive-MB-MBC (LA-MB-MBC) and Globally Adaptive-MB-MBC (GA-MB-MBC). Both methods monitor the concept drift over time using the average log-likelihood score and the Page-Hinkley test. Then, if a drift is detected, LA-MB-MBC adapts the current multi-dimensional Bayesian network classifier locally around each changed node, whereas GA-MB-MBC learns a new multi-dimensional Bayesian network classifier from scratch. Experimental study carried out using synthetic multi-dimensional data streams shows the merits of both proposed adaptive methods.

Simplifying data acquisition in plant canopies- Measurements of leaf angles with a cell phone

1. Canopies are complex multilayered structures comprising individual plant crowns exposing a multifaceted surface area to sunlight. Foliage arrangement and properties are the main mediators of canopy functions. The leaves act as light traps whose exposure to sunlight varies with time of the day, date and latitude in a trade-off between photosynthetic light harvesting and excessive or photoinhibitory light avoidance. To date, ecological research based upon leaf sampling has been limited by the available echnology, with which data acquisition becomes labour intensive and time-consuming, given the verwhelming number of leaves involved. 2. In the present study, our goal involved developing a tool capable of easuring a sufficient number of leaves to enable analysis of leaf populations, tree crowns and canopies.We specifically tested whether a cell phone working as a 3Dpointer could yield reliable, repeatable and valid leaf anglemeasurements with a simple gesture. We evaluated the accuracy of this method under controlled conditions, using a 3D digitizer, and we compared performance in the field with the methods commonly used. We presented an equation to estimate the potential proportion of the leaf exposed to direct sunlight (SAL) at any given time and compared the results with those obtained bymeans of a graphicalmethod. 3. We found a strong and highly significant correlation between the graphical methods and the equation presented. The calibration process showed a strong correlation between the results derived from the two methods with amean relative difference below 10%. Themean relative difference in calculation of instantaneous exposure was below 5%. Our device performed equally well in diverse locations, in which we characterized over 700 leaves in a single day. 4. The newmethod, involving the use of a cell phone, ismuchmore effective than the traditionalmethods or digitizers when the goal is to scale up from leaf position to performance of leaf populations, tree crowns or canopies. Our methodology constitutes an affordable and valuable tool within which to frame a wide range of ecological hypotheses and to support canopy modelling approaches.

How do various maize crop models vary in their responses to climate change factors?

Comments This article is a U.S. government work, and is not subject to copyright in the United States. Abstract Potential consequences of climate change on crop production can be studied using mechanistic crop simulation models. While a broad variety of maize simulation models exist, it is not known whether different models diverge on grain yield responses to changes in climatic factors, or whether they agree in their general trends related to phenology, growth, and yield. With the goal of analyzing the sensitivity of simulated yields to changes in temperature and atmospheric carbon dioxide concentrations [CO2], we present the largest maize crop model intercomparison to date, including 23 different models. These models were evaluated for four locations representing a wide range of maize production conditions in the world: Lusignan (France), Ames (USA), Rio Verde (Brazil) and Morogoro (Tanzania). While individual models differed considerably in absolute yield simulation at the four sites, an ensemble of a minimum number of models was able to simulate absolute yields accurately at the four sites even with low data for calibration, thus suggesting that using an ensemble of models has merit. Temperature increase had strong negative influence on modeled yield response of roughly 0.5 Mg ha 1 per °C. Doubling [CO2] from 360 to 720 lmol mol 1 increased grain yield by 7.5% on average across models and the sites. That would therefore make temperature the main factor altering maize yields at the end of this century. Furthermore, there was a large uncertainty in the yield response to [CO2] among models. Model responses to temperature and [CO2] did not differ whether models were simulated with low calibration information or, simulated with high level of calibration information.

Learning Bayesian networks from data by the incremental compilation of new network polynomials

Probabilistic graphical models are a huge research field in artificial intelligence nowadays. The scope of this work is the study of directed graphical models for the representation of discrete distributions. Two of the main research topics related to this area focus on performing inference over graphical models and on learning graphical models from data. Traditionally, the inference process and the learning process have been treated separately, but given that the learned models structure marks the inference complexity, this kind of strategies will sometimes produce very inefficient models. With the purpose of learning thinner models, in this master thesis we propose a new model for the representation of network polynomials, which we call polynomial trees. Polynomial trees are a complementary representation for Bayesian networks that allows an efficient evaluation of the inference complexity and provides a framework for exact inference. We also propose a set of methods for the incremental compilation of polynomial trees and an algorithm for learning polynomial trees from data using a greedy score+search method that includes the inference complexity as a penalization in the scoring function.