Towards a unified sentiment lexicon based on graphics processing units

This paper presents an approach to create what we have called a Unified Sentiment Lexicon (USL). This approach aims at aligning, unifying, and expanding the set of sentiment lexicons which are available on the web in order to increase their robustness of coverage. One problem related to the task of the automatic unification of different scores of sentiment lexicons is that there are multiple lexical entries for which the classification of positive, negative, or neutral {P, Z, N} depends on the unit of measurement used in the annotation methodology of the source sentiment lexicon. Our USL approach computes the unified strength of polarity of each lexical entry based on the Pearson correlation coefficient which measures how correlated lexical entries are with a value between 1 and -1, where 1 indicates that the lexical entries are perfectly correlated, 0 indicates no correlation, and -1 means they are perfectly inversely correlated and so is the UnifiedMetrics procedure for CPU and GPU, respectively. Another problem is the high processing time required for computing all the lexical entries in the unification task. Thus, the USL approach computes a subset of lexical entries in each of the 1344 GPU cores and uses parallel processing in order to unify 155802 lexical entries. The results of the analysis conducted using the USL approach show that the USL has 95.430 lexical entries, out of which there are 35.201 considered to be positive, 22.029 negative, and 38.200 neutral. Finally, the runtime was 10 minutes for 95.430 lexical entries; this allows a reduction of the time computing for the UnifiedMetrics by 3 times.

Hacia un Lexicón Unificado de Sentimientos basado en Unidades de Procesamiento Gráfico

Esta tesis presenta un modelo, una metodología, una arquitectura, varios algoritmos y programas para crear un lexicón de sentimientos unificado (LSU) que cubre cuatro lenguas: inglés, español, portugués y chino. El objetivo principal es alinear, unificar, y expandir el conjunto de lexicones de sentimientos disponibles en Internet y los desarrollados a lo largo de esta investigación. Así, el principal problema a resolver es la tarea de unificar de forma automatizada los diferentes lexicones de sentimientos obtenidos por el crawler CSR, porque la unidad de medida para asignar la intensidad de los valores de la polaridad (de forma manual, semiautomática y automática) varía de acuerdo con las diferentes metodologías utilizadas para la construcción de cada lexicón. La representación codificada de la estructura de datos de los términos presenta también una variación en la estructura de lexicón a lexicón. Por lo que al unificar en un lexicón de sentimientos se hace posible la reutilización del conocimiento recopilado por los diferentes grupos de investigación y se incrementa, a la vez, el alcance, la calidad y la robustez de los lexicones. Nuestra metodología LSU calcula un valor unificado de la intensidad de la polaridad para cada entrada léxica que está presente en al menos dos de los lexicones de sentimientos que forman parte de este estudio. En contraste, las entradas léxicas que no son comunes en al menos dos de los lexicones conservan su valor original. El coeficiente de Pearson resultante permite medir la correlación existente entre las entradas léxicas asignándoles un rango de valores de uno a menos uno, donde uno indica que los valores de los términos están perfectamente correlacionados, cero indica que no existe correlación y menos uno significa que están inversamente correlacionados. Este procedimiento se lleva acabo con la función de MetricasUnificadas tanto en la CPU como en la GPU. Otro problema a resolver es el tiempo de procesamiento que se requiere para realizar la tarea de unificación de la intensidad de la polaridad y con ello alcanzar una cobertura mayor de lemas en los lexicones de sentimientos existentes. Asimismo, la metodología LSU utiliza el procesamiento paralelo para unificar los 155 802 términos. El algoritmo LSU procesa mediante cargas iguales el subconjunto de entradas léxicas en cada uno de los 1344 núcleos en la GPU. Los resultados de nuestro análisis arrojaron un total de 95 430 entradas léxicas donde 35 201 obtuvieron valores positivos, 22 029 negativos y 38 200 neutrales. Finalmente, el tiempo de ejecución fue de 2,506 segundos para el total de las entradas léxicas, lo que permitió reducir el procesamiento de cómputo hasta en una tercera parte con respecto al algoritmo secuencial. De estos resultados se concluye que al lograr un lexicón de sentimientos unificado que permite homogeneizar la intensidad de la polaridad de las unidades léxicas (con valores positivos, negativos y neutrales) deriva no sólo en el análisis semántico del corpus basado en los términos con una mayor carga de polaridad, o del resumen de las valoraciones o las tendencias de neuromarketing, sino también en aplicaciones como el etiquetado subjetivo de sitios web o de portales sintácticos y semánticos, por mencionar algunas. ABSTRACT This thesis presents an approach to create what we have called a Unified Sentiment Lexicon (USL). This approach aims at aligning, unifying, and expanding the set of sentiment lexicons which are available on the web in order to increase their robustness of coverage. One problem related to the task of the automatic unification of different scores of sentiment lexicons is that there are multiple lexical entries for which the classification of positive, negative, or neutral P, N, Z depends on the unit of measurement used in the annotation methodology of the source sentiment lexicon. Our USL approach computes the unified strength of polarity of each lexical entry based on the Pearson correlation coefficient which measures how correlated lexical entries are with a value between 1 and - 1 , where 1 indicates that the lexical entries are perfectly correlated, 0 indicates no correlation, and -1 means they are perfectly inversely correlated and so is the UnifiedMetrics procedure for CPU and GPU, respectively. Another problem is the high processing time required for computing all the lexical entries in the unification task. Thus, the USL approach computes a subset of lexical entries in each of the 1344 GPU cores and uses parallel processing in order to unify 155,802 lexical entries. The results of the analysis conducted using the USL approach show that the USL has 95,430 lexical entries, out of which there are 35,201 considered to be positive, 22,029 negative, and 38,200 neutral. Finally, the runtime was 2.505 seconds for 95,430 lexical entries; this allows a reduction of the time computing for the UnifiedMetrics by 3 times with respect to the sequential implementation. A key contribution of this work is that we preserve the use of a unified sentiment lexicon for all tasks. Such lexicon is used to define resources and resource-related properties that can be verified based on the results of the analysis and is powerful, general and extensible enough to express a large class of interesting properties. Some applications of this work include merging, aligning, pruning and extending the current sentiment lexicons.

Análisis, desarrollo y optimización de un sistema para el diseño de redes topográficas valorando el modelo digital del terreno

El presente trabajo tiene como objetivo general el análisis de las técnicas de diseño y optimización de redes topográficas, observadas mediante topografía convencional (no satelital) el desarrollo e implementación de un sistema informático capaz de ayudar a la definición de la geometría más fiable y precisa, en función de la orografía del terreno donde se tenga que ubicar. En primer lugar se realizará un estudio de la metodología del ajuste mediante mínimos cuadrados y la propagación de varianzas, para posteriormente analizar su dependencia de la geometría que adopte la red. Será imprescindible determinar la independencia de la matriz de redundancia (R) de las observaciones y su total dependencia de la geometría, así como la influencia de su diagonal principal (rii), números de redundancia, para garantizar la máxima fiabilidad interna de la misma. También se analizará el comportamiento de los números de redundancia (rii) en el diseño de una red topográfica, la variación de dichos valores en función de la geometría, analizando su independencia respecto de las observaciones así como los diferentes niveles de diseño en función de los parámetros y datos conocidos. Ha de señalarse que la optimización de la red, con arreglo a los criterios expuestos, está sujeta a los condicionantes que impone la necesidad de que los vértices sean accesibles, y además sean visibles entre sí, aquellos relacionados por observaciones, situaciones que dependen esencialmente del relieve del terreno y de los obstáculos naturales o artificiales que puedan existir. Esto implica la necesidad de incluir en el análisis y en el diseño, cuando menos de un modelo digital del terreno (MDT), aunque lo más útil sería la inclusión en el estudio del modelo digital de superficie (MDS), pero esta opción no siempre será posible. Aunque el tratamiento del diseño esté basado en un sistema bidimensional se estudiará la posibilidad de incorporar un modelo digital de superficie (MDS); esto permitirá a la hora de diseñar el emplazamiento de los vértices de la red la viabilidad de las observaciones en función de la orografía y los elementos, tanto naturales como artificiales, que sobre ella estén ubicados. Este sistema proporcionaría, en un principio, un diseño óptimo de una red constreñida, atendiendo a la fiabilidad interna y a la precisión final de sus vértices, teniendo en cuenta la orografía, lo que equivaldría a resolver un planteamiento de diseño en dos dimensiones y media1; siempre y cuando se dispusiera de un modelo digital de superficie o del terreno. Dado que la disponibilidad de obtener de manera libre el MDS de las zonas de interés del proyecto, hoy en día es costoso2, se planteará la posibilidad de conjuntar, para el estudio del diseño de la red, de un modelo digital del terreno. Las actividades a desarrollar en el trabajo de esta tesis se describen en esta memoria y se enmarcan dentro de la investigación para la que se plantean los siguientes objetivos globales: 1. Establecer un modelo matemático del proceso de observación de una red topográfica, atendiendo a todos los factores que intervienen en el mismo y a su influencia sobre las estimaciones de las incógnitas que se obtienen como resultado del ajuste de las observaciones. 2. Desarrollar un sistema que permita optimizar una red topográfica en sus resultados, aplicando técnicas de diseño y simulación sobre el modelo anterior. 3. Presentar una formulación explícita y rigurosa de los parámetros que valoran la fiabilidad de una red topográfica y de sus relaciones con el diseño de la misma. El logro de este objetivo se basa, además de en la búsqueda y revisión de las fuentes, en una intensa labor de unificación de notaciones y de construcción de pasos intermedios en los desarrollos matemáticos. 4. Elaborar una visión conjunta de la influencia del diseño de una red, en los seis siguientes factores (precisiones a posteriori, fiabilidad de las observaciones, naturaleza y viabilidad de las mismas, instrumental y metodología de estacionamiento) como criterios de optimización, con la finalidad de enmarcar el tema concreto que aquí se aborda. 5. Elaborar y programar los algoritmos necesarios para poder desarrollar una aplicación que sea capaz de contemplar las variables planteadas en el apartado anterior en el problema del diseño y simulación de redes topográficas, contemplando el modelo digital de superficie. Podrían considerarse como objetivos secundarios, los siguientes apartados: Desarrollar los algoritmos necesarios para interrelacionar el modelo digital del terreno con los propios del diseño. Implementar en la aplicación informática la posibilidad de variación, por parte del usuario, de los criterios de cobertura de los parámetros (distribución normal o t de Student), así como los grados de fiabilidad de los mismos ABSTRACT The overall purpose of this work is the analysis of the techniques of design and optimization for geodetic networks, measured with conventional survey methods (not satellite), the development and implementation of a computational system capable to help on the definition of the most liable and accurate geometry, depending on the land orography where the network has to be located. First of all, a study of the methodology by least squares adjustment and propagation of variances will be held; then, subsequently, analyze its dependency of the geometry that the network will take. It will be essential to determine the independency of redundancy matrix (R) from the observations and its absolute dependency from the network geometry, as well as the influence of the diagonal terms of the R matrix (rii), redundancy numbers, in order to ensure maximum re liability of the network. It will also be analyzed first the behavior of redundancy numbers (rii) in surveying network design, then the variation of these values depending on the geometry with the analysis of its independency from the observations, and finally the different design levels depending on parameters and known data. It should be stated that network optimization, according to exposed criteria, is subject to the accessibility of the network points. In addition, common visibility among network points, which of them are connected with observations, has to be considered. All these situations depends essentially on the terrain relief and the natural or artificial obstacles that should exist. Therefore, it is necessary to include, at least, a digital terrain model (DTM), and better a digital surface model (DSM), not always available. Although design treatment is based on a bidimensional system, the possibility of incorporating a digital surface model (DSM) will be studied; this will allow evaluating the observations feasibility based on the terrain and the elements, both natural and artificial, which are located on it, when selecting network point locations. This system would provide, at first, an optimal design of a constrained network, considering both the internal reliability and the accuracy of its points (including the relief). This approach would amount to solving a “two and a half dimensional”3 design, if a digital surface model is available. As the availability of free DSM4 of the areas of interest of the project today is expensive, the possibility of combining a digital terrain model will arise. The activities to be developed on this PhD thesis are described in this document and are part of the research for which the following overall objectives are posed: 1. To establish a mathematical model for the process of observation of a survey network, considering all the factors involved and its influence on the estimates of the unknowns that are obtained as a result of the observations adjustment. 2. To develop a system to optimize a survey network results, applying design and simulation techniques on the previous model. 3. To present an explicit and rigorous formulation of parameters which assess the reliability of a survey network and its relations with the design. The achievement of this objective is based, besides on the search and review of sources, in an intense work of unification of notation and construction of intermediate steps in the mathematical developments. 4. To develop an overview of the influence on the network design of six major factors (posterior accuracy, observations reliability, viability of observations, instruments and station methodology) as optimization criteria, in order to define the subject approached on this document. 5. To elaborate and program the algorithms needed to develop an application software capable of considering the variables proposed in the previous section, on the problem of design and simulation of surveying networks, considering the digital surface model. It could be considered as secondary objectives, the following paragraphs: To develop the necessary algorithms to interrelate the digital terrain model with the design ones. To implement in the software application the possibility of variation of the coverage criteria parameters (normal distribution or Student t test) and therefore its degree of reliability.

Differential invariants of second-order ordinary differential equations

The notion of a differential invariant for systems of second-order differential equations on a manifold M with respect to the group of vertical automorphisms of the projection is de?ned and the Chern connection attached to a SODE allows one to determine a basis for second-order differential invariants of a SODE.

Independent AND-parallel implementation of narrowing

We present a parallel graph narrowing machine, which is used to implement a functional logic language on a shared memory multiprocessor. It is an extensión of an abstract machine for a purely functional language. The result is a programmed graph reduction machine which integrates the mechanisms of unification, backtracking, and independent and-parallelism. In the machine, the subexpressions of an expression can run in parallel. In the case of backtracking, the structure of an expression is used to avoid the reevaluation of subexpressions as far as possible. Deterministic computations are detected. Their results are maintained and need not be reevaluated after backtracking.

Combined determination of sharing and freeness of program variables through abstract interpretation

In this paper, abstract interpretation algorithms are described for computing the sharmg as well as the freeness information about the run-time instantiations of program variables. An abstract domain is proposed which accurately and concisely represents combined freeness and sharing information for program variables. Abstract unification and all other domain-specific functions for an abstract interpreter working on this domain are presented. These functions are illustrated with an example. The importance of inferring freeness is stressed by showing (1) the central role it plays in non-strict goal independence, and (2) the improved accuracy it brings to the analysis of sharing information when both are computed together. Conversely, it is shown that keeping accurate track of sharing allows more precise inference of freeness, thus resulting in an overall much more powerful abstract interpreter.

Effectiveness of combined sharing and freeness analysis using abstract interpretation

This paper presents improved unification algorithms, an implementation, and an analysis of the effectiveness of an abstract interpreter based on the sharing + freeness domain presented in a previous paper, which was designed to accurately and concisely represent combined freeness and sharing information for program variables. We first briefly review this domain and the unification algorithms previously proposed. We then improve these algorithms and correct them to deal with some cases which were not well analyzed previously, illustrating the improvement with an example. We then present the implementation of the improved algorithm and evaluate its performance by comparing the effectiveness of the information inferred to that of other interpreters available to us for an application (program parallelization) that is common to all these interpreters. All these systems have been embedded in a real parallelizing compiler. Effectiveness of the analysis is measured in terms of actual final performance of the system: i.e. in terms of the actual speedups obtained. The results show good performance for the combined domain in that it improves the accuracy of both types of information and also in that the analyzer using the combined domain is more effective in the application than any of the other analyzers it is compared to.

Experiences of spanish public administration in requirements management and acquisition management processes

This paper shows the main contributions of the 1st Symposium on Improvement Process Models and Software Quality of Public Administrations. The obtained results expose the need to promote the implementation of Software Maturity Models and show possible advantages of its application in software processes of Public Administrations. Specifically, it was analyzed the current status in two process areas: Requirements Management and Subcontracting Management.

Flammability properties of dry sewage sludges

Dry sewage sludge are being considered as a possible energy source for direct firing. They have interesting properties to be used as an alternative fuel, but also other characteristics must be considered from the point of view of its safe operation: the most important are ignition sensitivity, explosion severity, thermal sensitivity and thermal stability. The aim of this study was to determine if sewage sludge have different characteristics due to different locations or seasons and how this influences their flammability properties. To study these characteristics sludge samples were selected from different locations in Spain, taken during different seasons. In addition, relationships between flammability parameters and chemical analysis were observed. Some parameters can be controlled during normal operation, such as granulometry or humidity, and may mean a decrease in the risk of explosion. Those relationships are well known for other dusts materials, like coal, but not yet for sewage sludge dusts. Finally, properties related to spontaneous combustion were determined (thermal susceptibility and stability). The addition of those properties to the study provides an overview of the thermal behavior of sewage sludge during their utilization, including transport and storage.

A Knowledge-based Approach for Automatic Generation of Summaries of Behavior

Effective automatic summarization usually requires simulating human reasoning such as abstraction or relevance reasoning. In this paper we describe a solution for this type of reasoning in the particular case of surveillance of the behavior of a dynamic system using sensor data. The paper first presents the approach describing the required type of knowledge with a possible representation. This includes knowledge about the system structure, behavior, interpretation and saliency. Then, the paper shows the inference algorithm to produce a summarization tree based on the exploitation of the physical characteristics of the system. The paper illustrates how the method is used in the context of automatic generation of summaries of behavior in an application for basin surveillance in the presence of river floods.

Benefits of TSPi in a software project under a small settings enviroment

This article presents a case study about the TSPi benefits in a software project under a Small Settings environment. An adapted process based on the TSPi was defined. The pilot project had a schedule and budget restricted. The process began collecting historical projects data in order to get a measure repository. The project was launched defining the following goals: increase the productivity, reduce the test time and improve the product quality. Finally, the results were analysed and the goals were verified.