Desarrollo de una herramienta para analizar métodos de reconocimiento de emociones

Desde hace más de 20 años, muchos grupos de investigación trabajan en el estudio de técnicas de reconocimiento automático de expresiones faciales. En los últimos años, gracias al avance de las metodologías, ha habido numerosos avances que hacen posible una rápida detección de las caras presentes en una imagen y proporcionan algoritmos de clasificación de expresiones. En este proyecto se realiza un estudio sobre el estado del arte en reconocimiento automático de emociones, para conocer los diversos métodos que existen en el análisis facial y en el reconocimiento de la emoción. Con el fin de poder comparar estos métodos y otros futuros, se implementa una herramienta modular y ampliable y que además integra un método de extracción de características que consiste en la obtención de puntos de interés en la cara y dos métodos para clasificar la expresión, uno mediante comparación de desplazamientos de los puntos faciales, y otro mediante detección de movimientos específicos llamados unidades de acción. Para el entrenamiento del sistema y la posterior evaluación del mismo, se emplean las bases de datos Cohn-Kanade+ y JAFFE, de libre acceso a la comunidad científica. Después, una evaluación de estos métodos es llevada a cabo usando diferentes parámetros, bases de datos y variando el número de emociones. Finalmente, se extraen conclusiones del trabajo y su evaluación, proponiendo las mejoras necesarias e investigación futura. ABSTRACT. Currently, many research teams focus on the study of techniques for automatic facial expression recognition. Due to the appearance of digital image processing, in recent years there have been many advances in the field of face detection, feature extraction and expression classification. In this project, a study of the state of the art on automatic emotion recognition is performed to know the different methods existing in facial feature extraction and emotion recognition. To compare these methods, a user friendly tool is implemented. Besides, a feature extraction method is developed which consists in obtaining 19 facial feature points. Those are passed to two expression classifier methods, one based on point displacements, and one based on the recognition of facial Action Units. Cohn-Kanade+ and JAFFE databases, both freely available to the scientific community, are used for system training and evaluation. Then, an evaluation of the methods is performed with different parameters, databases and varying the number of emotions. Finally, conclusions of the work and its evaluation are extracted, proposing some necessary improvements and future research.

Methodological Proposal for Teamwork Evaluation in the field of training Project Management.

Teamwork, is one of the abilities that today is highly valued in the professional arena with a great importance for various personal and interpersonal skills associated with it. In this context, the Technical University of Madrid, is developing a coordinated educational innovation project, which main objective is to develop methodological and assessment tools for the acquisition of personal skills necessary to improve the employability of graduates and their skills for project management. Within this context, this paper proposes a methodology composed of various activities and indicators, as well as specific assessment instruments linked to the teamwork competence. Through a series of systematic steps it was allowed the design of an instrument and construction of a scale for measuring the competence of teamwork. The practical application of the methodology has been carried out in Projects lectures from different Schools of Engineering at the Technical University of Madrid, which results are presented in this document as a pilot experience. Results show the various aspects and methods that teachers should consider in evaluating the competence of the work, including analysis of the quality of results, through reliability and construct validity. On the other hand, show the advantages of applying this methodology in the field of project management teaching.

Evaluation of a PET Prototype Using LYSO:Ce Monolithic Detector Block

We have analyzed the performance of a PET demonstrator formed by two sectors of four monolithic detector blocks placed face-to-face. Both front-end and read-out electronics have been evaluated by means of coincidence measurements using a rotating 22Na source placed at the center of the sectors in order to emulate the behavior of a complete full ring. A continuous training method based on neural network (NN) algorithms has been carried out to determine the entrance points over the surface of the detectors. Reconstructed images from 1 MBq 22Na point source and 22Na Derenzo phantom have been obtained using both filtered back projection (FBP) analytic methods and the OSEM 3D iterative algorithm available in the STIR software package [1]. Preliminary data on image reconstruction from a 22Na point source with Ø = 0.25 mm show spatial resolutions from 1.7 to 2.1 mm FWHM in the transverse plane. The results confirm the viability of this design for the development of a full-ring brain PET scanner compatible with magnetic resonance imaging for human studies.

Training project management complexity in postgraduate and continuing education programs: a learning strategy in the eshe (European Space of Higher Education) framework

The objective of this paper is to address the methodological process of a teaching strategy for training project management complexity in postgraduate programs. The proposal is made up of different methods —intuitive, comparative, deductive, case study, problem-solving Project-Based Learning— and different activities inside and outside the classroom. This integration of methods motivated the current use of the concept of “learning strategy”. The strategy has two phases: firstly, the integration of the competences —technical, behavioral and contextual—in real projects; and secondly, the learning activity was oriented in upper level of knowledge, the evaluating the complexity for projects management in real situations. Both the competences in the learning strategy and the Project Complexity Evaluation are based on the ICB of IPMA. The learning strategy is applied in an international Postgraduate Program —Erasmus Mundus Master of Science— with the participation of five Universities of the European Union. This master program is fruit of a cooperative experience from one Educative Innovation Group of the UPM -GIE-Project-, two Research Groups of the UPM and the collaboration with other external agents to the university. Some reflections on the experience and the main success factors in the learning strategy were presented in the paper

Training project management complexity in postgraduate and continuing education programs: a learning strategy in the eshe (European Space of Higher Education) framework

The objective of this paper is to address the methodological process of a teaching strategy for training project management complexity in postgraduate programs. The proposal is made up of different methods —intuitive, comparative, deductive, case study, problem-solving Project-Based Learning— and different activities inside and outside the classroom. This integration of methods motivated the current use of the concept of ―learning strategy‖. The strategy has two phases: firstly, the integration of the competences —technical, behavioral and contextual—in real projects; and secondly, the learning activity was oriented in upper level of knowledge, the evaluating the complexity for projects management in real situations. Both the competences in the learning strategy and the Project Complexity Evaluation are based on the ICB of IPMA. The learning strategy is applied in an international Postgraduate Program —Erasmus Mundus Master of Science— with the participation of five Universities of the European Union. This master program is fruit of a cooperative experience from one Educative Innovation Group of the UPM -GIE-Project-, two Research Groups of the UPM and the collaboration with other external agents to the university. Some reflections on the experience and the main success factors in the learning strategy were presented in the paper.

Patrol team language identification system for DARPA RATS P1 evaluation

This paper describes the language identification (LID) system developed by the Patrol team for the first phase of the DARPA RATS (Robust Automatic Transcription of Speech) program, which seeks to advance state of the art detection capabilities on audio from highly degraded communication channels. We show that techniques originally developed for LID on telephone speech (e.g., for the NIST language recognition evaluations) remain effective on the noisy RATS data, provided that careful consideration is applied when designing the training and development sets. In addition, we show significant improvements from the use of Wiener filtering, neural network based and language dependent i-vector modeling, and fusion.

Systems and technologies for objective evaluation of technical skills in laparoscopic surgery

Minimally invasive surgery is a highly demanding surgical approach regarding technical requirements for the surgeon, who must be trained in order to perform a safe surgical intervention. Traditional surgical education in minimally invasive surgery is commonly based on subjective criteria to quantify and evaluate surgical abilities, which could be potentially unsafe for the patient. Authors, surgeons and associations are increasingly demanding the development of more objective assessment tools that can accredit surgeons as technically competent. This paper describes the state of the art in objective assessment methods of surgical skills. It gives an overview on assessment systems based on structured checklists and rating scales, surgical simulators, and instrument motion analysis. As a future work, an objective and automatic assessment method of surgical skills should be standardized as a means towards proficiency-based curricula for training in laparoscopic surgery and its certification.

The influence of transparency on self-evaluation as part of the university accreditation process in Peru.

In order to show the choice of transparency as the guiding principle of the accreditation process, the article evaluates its influence on the fundamental subprocess of self-evaluation, thereby confirming that transparency is an essential tool for continuous improvement of academic processes and those of educational quality management. It fosters educational innovation and permits the sustainability of the continuous accreditation process over time, resulting in greater probabilities of university self-regulation through systemization of the process, with the objective of continuous improvement of university degree programs. The article analyzes the influence of transparency on each activity of the self-evaluation process according to the Peruvian accreditation model prepared under the total quality approach, as a reference for other accreditation models, proposing concrete transparency actions and evaluating its influence on the stakeholder groups in the self-evaluation process, as well as on the efficiency and effectiveness of the process. It is concluded that transparency has a positive influence on the training of human capital and the formation of the university?s organizational culture, facilitating dissemination, understanding and involvement of the stakeholder groups in the continuous improvement of accreditation activities and increasing their acceptance of change and commitment to the process. It is confirmed that transparency contributes toward increasing the efficiency index of the self-evaluation process by reducing operating costs through adequate, accessible, timely contribution of information by the stakeholders and through the optimization of the time spent gathering relevant information. In addition, it is concluded that transparency contributes toward increasing the effectiveness index of self-evaluation by facilitating the achievement of its objectives through synthetic, useful, reliable interpretation of the education situation and the formulation of feasible improvement plans based on the adequacy, relevance, visibility, pertinence and truthfulness of the information analyzed.

Biomechanical evaluation of the singing voice

Teaching the adequate use of the singing voice conveys a lot of knowledge in musical performance as well as in objective estimation techniques involving the use of air, muscles, room and body acoustics, and the tuning of a fine instrument as the human voice. Although subjective evaluation and training is a very delicate task to be carried out only by expert singers, biomedical engineering may help contributing with well-funded methodologies developed for the study of voice pathology. The present work is a preliminary study of exploratory character describing the performance of a student singer in a regular classroom under the point of view of vocal fold biomechanics. Estimates of biomechanical parameters obtained from singing voice are given and their potential use is discussed.

Global research alliance modelling platform (GRAMP): an open web platform for modelling greenhouse gas emissions from agro-ecosystems

Carbon (C) and nitrogen (N) process-based models are important tools for estimating and reporting greenhouse gas emissions and changes in soil C stocks. There is a need for continuous evaluation, development and adaptation of these models to improve scientific understanding, national inventories and assessment of mitigation options across the world. To date, much of the information needed to describe different processes like transpiration, photosynthesis, plant growth and maintenance, above and below ground carbon dynamics, decomposition and nitrogen mineralization. In ecosystem models remains inaccessible to the wider community, being stored within model computer source code, or held internally by modelling teams. Here we describe the Global Research Alliance Modelling Platform (GRAMP), a web-based modelling platform to link researchers with appropriate datasets, models and training material. It will provide access to model source code and an interactive platform for researchers to form a consensus on existing methods, and to synthesize new ideas, which will help to advance progress in this area. The platform will eventually support a variety of models, but to trial the platform and test the architecture and functionality, it was piloted with variants of the DNDC model. The intention is to form a worldwide collaborative network (a virtual laboratory) via an interactive website with access to models and best practice guidelines; appropriate datasets for testing, calibrating and evaluating models; on-line tutorials and links to modelling and data provider research groups, and their associated publications. A graphical user interface has been designed to view the model development tree and access all of the above functions.

Evaluation of the competitiveness of agri-food sector in the región of the Alcarria Conquense(Spain)

The agrifood industry, like other sectors, faces the ongoing challenge of improving their competitiveness in order to strengthen its market presence and cater to the growing global population. This research measures the competitiveness of the agrifood industry in the region of La Alcarria Conquense (Spain), in the framework of the evaluation of programs in the territory that have aimed at improving and enhancing this sector. Through building the competitiveness profiles (Porter, 1990) and cluster analysis we have identified six competitive strategy patterns in food companies in the region. In addition, we have analyzed each of the areas of competitiveness and we can identify the strengths and weaknesses of the sector, and identify recommendations for increasing the responsiveness of the territory. Among the defining characteristics are the lack of association, the limitation on payment systems or virtual absence of training and innovation. However, programs to support the sector are highly valued and reverse in the long-term viability of these companies.

Novel educational assessment for building structures: Automatic evaluation of on-line graphics

This article presents a new automatic evaluation for on-line graphics, its application and the numerous advantages achieved applying this developed correcting method. The software application developed by the Innovation in Education Group “E4”, from the Technical University of Madrid, is oriented for the online self-assessment of the graphic drawings that students carry out as continuous training. The adaptation to the European Higher Educational Area is an important opportunity to research about the possibilities of on-line education assessment. In this way, a new software tool has been developed for continuous self-testing by undergraduates. Using this software it is possible to evaluate the graphical answer of the students. Thus, the drawings made on-line by students are automatically corrected according to the geometry (straight lines, sloping lines or second order curves) and by sizes (depending on the specific values which define the graphics).

Evaluación del rendimiento de jugadores y equipos de fútbol de nivel de élite en diferentes contextos situacionales = Evaluation on Match Performances of Professional Football Players and Teams under Different Situational Conditions = 不同情境条件下高水平足球运动员和运动队的比赛表现分析

El análisis del rendimiento en deportes juega un papel esencial en el fútbol profesional. Aunque el estudio del análisis del juego en fútbol se ha utilizado desde diferentes ámbitos y situaciones, todavía existen diferentes aspectos y componentes del juego que siguen sin estar estudiados. En este sentido existen diferentes aspectos que deben de superar los estudios previos centrados en el componente descriptivo tales como el uso de variables/ indicadores de rendimiento que no se han definido ni estudiado, la validez de los métodos observaciones que no han sido testados con los softwares específicos en fútbol, la aplicación y utilidad de los resultados, así como las limitaciones del estudio de las variables situacionales/contextuales. Con el objetivo de cubrir las citadas limitaciones se han diseñado 6 estudios independientes e inter-relacionados que tratan de estudiar los aspectos anteriormente referidos. El primer estudio evalua la fiabilidad inter-observadores de las estadísticas de juego de la empresa privada OPTA Sportsdata, estos datos son la muestra de estudio de la presente tesis doctoral. Dos grupos de observadores experimentados se requieren para analizar un partido de la liga española de manera independiente. Los resultados muestran que los eventos de equipos y porteros codificados por los inter-operadores alcanzan un acuerdo muy bueno (valores kappa entre 0.86 y 0.94). La validez inter-observadores de las acciones de juego y los datos de jugadores individuales se evaluó con elevados niveles de acuerdo (valores del coeficiente de correlación intraclase entre 0.88 hasta 1.00, el error típico estandarizado variaba entre 0.00 hasta 0.37). Los resultados sugieren que las estadísticas de juego registradas por los operadores de la empresa OPTA Sportsdata están bien entrenados y son fiables. El segundo, tercer y cuarto estudio se centran en resaltar la aplicabilidad del análisis de rendimiento en el fútbol así como para explicar en profundidad las influencias de las variables situacionales. Utilizando la técnica de los perfiles de rendimiento de jugadores y equipos de fútbol se puede evaluar y comparar de manera gráfica, fácil y visual. Así mismo, mediante esta técnica se puede controlar el efecto de las variables situacionales (localización del partido, nivel del equipo y del oponente, y el resultado final del partido). Los perfiles de rendimiento de porteros (n = 46 porteros, 744 observaciones) y jugadores de campo (n = 409 jugadores, 5288 observaciones) de la primera division professional de fútbol Española (La Liga, temporada 2012-13), los equipos (n = 496 partidos, 992 observaciones) de la UEFA Champions League (temporadas 2009-10 a 2012-13) fueron analizados registrando la media, desviación típica, mediana, cuartiles superior e inferior y el recuento de valores de cada indicador de rendimiento y evento, los cuales se presentaron en su forma tipificada y normalizada. Los valores medios de los porteros de los equipos de diferentes niveles de La Liga y de los equipos de diferente nivel de la UEFA Champions League cuando jugaban en diferentes contextos de juego y situaciones (variables situacionales) fueron comparados utilizando el ANOVA de un factor y la prueba t para muestras independientes (localización del partido, diferencias entre casa y fuera), y fueron establecidos en los perfiles de red después de unificar todos los registros en la misma escala derivada con valores estandarizados. Mientras que las diferencias de rendimiento entre los jugadores de los mejores equipos (Top3) y los peores (Bottom3) fueron comparados mediante el uso de diferencias en la magnitud del tamaño del efecto. El quinto y el sexto estudio analizaban el rendimiento del fútbol desde un punto de vista de predicción del rendimiento. El modelo linear general y el modelo lineal general mixto fue empleado para analizar la magnitud de las relaciones de los indicadores y estadísticas de juego con el resultado final del partido en función del tipo de partido (partidos ajustados o todos los partidos) en la fase de grupos de la Copa del Mundo 2014 de Brasil (n = 48 partidos, 38 partidos ajustados) y La Liga 2012-13 (n = 320 partidos ajustados). Las relaciones fueron evaluadas mediante las inferencias en la magnitud de las diferencias y se expresaron como partidos extra ganados o perdidos por cada 10 partidos mediante la variable calculada en 2 desviaciones típicas. Los resultados mostraron que, para los 48 partidos de la fase de grupos de la Copa del Mundo 2014, nueve variables tuvieron un efecto positive en la probabilidad de ganar (tiros, tiros a puerta, tiros de contraataque, tiros dentro del área, posesión de balón, pases en corto, media de secuencia de pases, duelos aéreos y entradas), cuatro tuvieron efectos negativos (tiros bloqueados, centros, regates y tarjetas amarillas), y otras 12 variables tenían efectos triviales o poco claros. Mientras que los 38 partidos ajustados, el efecto de duelos aéreos y tarjetas amarillas fueron triviales y claramente negativos respectivamente. En la La Liga, existió un efecto moderado positive para cada equipo para los tiros a puerta (3.4 victorias extras por cada 10 partidos; 99% IC ±1.0), y un efecto positivo reducido para tiros totales (1.7 victorias extrsa; ±1.0). Los efectos de la mayoría de los eventos se han relacionado con la posesión del balón, la cual obtuvo efectos negativos entre equipos (1.2 derrotas extras; ±1.0) pero un efecto positivo pequeño entra equipos (1.7 victorias extras; ±1.4). La localización del partido mostró un efecto positive reducido dentro de los equipos (1.9 victorias extras; ±0.9). Los resultados obtenidos en los perfiles y el modelado del rendimiento permiten ofrecer una información detallada y avanzada para el entrenamiento, la preparación previa a los partidos, el control de la competición y el análisis post-partido, así como la evaluación e identificación del talento de los jugadores. ABSTRACT Match performance analysis plays an important role in the modern professional football. Although the research in football match analysis is well-developed, there are still some issues and problems remaining in this field, which mainly include the lack of operational definitions of variables, reliability issues, applicability of the findings, the lack of contextual/situational variables, and focusing too much on descriptive and comparative analysis. In order to address these issues, six independent but related studies were conducted in the current thesis. The first study evaluated the inter-operator reliability of football match statistics from OPTA Sportsdata Company which is the data resourse of the thesis. Two groups of experienced operators were required to analyse a Spanish league match independently in the experiment. Results showed that team events and goalkeeper actions coded by independent operators reached a very good agreement (kappa values between 0.86 and 0.94). The inter-operator reliability of match actions and events of individual outfield players was also tested to be at a high level (intra-class correlation coefficients ranged from 0.88 to 1.00, standardised typical error varied from 0.00 to 0.37). These results suggest that the football match statistics collected by well-trained operators from OPTA Sportsdata Company are reliable. The second, third and fourth study aims to enhance the applicability of football match performance analysis and to explore deeply the influences of situational variables. By using a profiling technique, technical and tactical performances of football players and teams can be interpreted, evaluated and compared more easily and straightforwardly, meanwhile, influences and effects from situational variables (match location, strength of team and opposition, and match outcome) on the performances can be properly incorporated. Performance profiles of goalkeepers (n = 46 goalkeepers, 744 full match observations) and outfield players (n = 409 players, 5288 full match observations) from the Spanish First Division Professional Football League (La Liga, season 2012-13), teams (n = 496 matches, 992 observations) from UEFA Champions League (seasons 2009-10 to 2012-13) were set up by presenting the mean, standard deviation, median, lower and upper quartiles of the count values of each performance-related match action and event to represent their typical performances and spreads. Means of goalkeeper from different levels of team in La Liga and teams of different strength in UEFA Champions League when playing under different situational conditions were compared by using one-way ANOVA and independent sample t test (for match location, home and away differences), and were plotted into the same radar charts after unifying all the event counts by standardised score. While differences between the performances of outfield players from Top3 and from Bottom3 teams were compared by magnitude-based inferences. The fifth and sixth study aims to move from the descriptive and comparative football match analysis to a more predictive one. Generalised linear modelling and generalised mixed linear modelling were undertaken to quantify relationships of the performance-related match events, actions and variables with the match outcome in different types of games (close games and all games) in the group stage of 2014 Brazil FIFA World Cup (n = 48 games, 38 close games) and La Liga 2012-13 (n = 320 close games). Relationships were evaluated with magnitude-based inferences and were expressed as extra matches won or lost per 10 matches for an increase of two standard deviations of a variable. Results showed that, for all the 48 games in the group stage of 2014 FIFA World Cup, nine variables had clearly positive effects on the probability of winning (shot, shot on target, shot from counter attack, shot from inside area, ball possession, short pass, average pass streak, aerial advantage, and tackle), four had clearly negative effects (shot blocked, cross, dribble and red card), other 12 variabless had either trivial or unclear effects. While for the 38 close games, the effects of aerial advantage and yellow card turned to trivial and clearly negative, respectively. In the La Liga, there was a moderate positive within-team effect from shots on target (3.4 extra wins per 10 matches; 99% confidence limits ±1.0), and a small positive within-team effect from total shots (1.7 extra wins; ±1.0). Effects of most other match events were related to ball possession, which had a small negative within-team effect (1.2 extra losses; ±1.0) but a small positive between-team effect (1.7 extra wins; ±1.4). Game location showed a small positive within-team effect (1.9 extra wins; ±0.9). Results from the established performance profiles and modelling can provide detailed and straightforward information for training, pre-match preparations, in-match tactical approaches and post-match evaluations, as well as for player identification and development. 摘要 比赛表现分析在现代足球中起着举足轻重的作用。尽管如今对足球比赛表现分析的研究已经相对完善，但仍有很多不足之处。这些不足主要体现在：研究中缺乏对研究变量的清晰定义、数据信效度缺失、研究结果的实用性受限、比赛情境因素缺失以及过于集中在描述性和对比性分析等。针对这些问题，本论文通过六个独立而又相互联系的研究，进一步对足球比赛表现分析进行完善。 第一个研究对本论文的数据源--OPTA Sportsdata公司的足球比赛数据的信效度进行了实验检验。实验中，两组数据收集人员被要求对同一场西班牙足球甲级联赛的比赛进行分析。研究结果显示，两组收集人员记录下的球队比赛事件和守门员比赛行为具有高度的一致性(卡帕系数介于0.86和0.94)。收集人员输出的外场球员的比赛行为和比赛事件也具有很高的组间一致性(ICC相关系数介于0.88和1.00，标准化典型误差介于0.00和0.37)。实验结果证明了OPTA Sportsdata公司收集的足球比赛数据具有足够高的信效度。 第二、三、四个研究旨在提升足球比赛表现分析研究结果的实用性以及深度探讨比赛情境因素对足球比赛表现的影响。通过对足球运动员和运动队的比赛技战术表现进行档案创建，可以对运动员和运动队的比赛表现进行简直接而直观的呈现、评价和对比，同时，情境变量(比赛场地、球队和对手实力、比赛结果)对比赛表现的影响也可以被整合到表现档案中。本部分对2012-13赛季西班牙足球甲级联赛的参赛守门员(n = 46球员人次，744比赛场次)和外场球员(n = 409球员人次, 5288比赛场次)以及2009-10至2012-13赛季欧洲足球冠军联赛的参赛球队(n = 496比赛场次)的比赛技战术表现进行了档案创建。在表现档案中，各项比赛技战术指标的均值、标准差、中位数和大小四分位数被用来展现守门员、外场球员和球队的普遍表现和表现浮动性。方差分析(ANOVA)被用来对西甲不同水平球队的守门员、欧冠中不同水平球队在不同比赛情境下的普遍表现(各项指标的均值)进行对比，独立样本t检验被用来对比主客场比赛普遍表现的差异。数据量级推断(magnitude-based inferences)的方法则被用来对西甲前三名和最后三名球队外场球员的普遍表现进行对比分析。所有来自不同水平球队的运动员和不同水平运动队的各项比赛指标皆被转换成了标准分数，从而能把他们在各种不同比赛情境下的普遍表现(各项比赛指标的均值)投到相同的雷达图中进行直观的对比。 第五和第六个研究目的在于进行预测性足球比赛表现分析，从而跨越之前固有的描述性和对比性分析。广义线性模型和广义混合线性模型被用来对2014年巴西世界杯小组赛(n = 48 比赛场次，38小分差场次)和2012-13赛季西甲联赛(n = 320小分差场次)的比赛中各表现相关比赛事件、行为和变量与比赛结果(胜、平、负)的关系进行建模。模型中的关系通过数据量级推断(magnitude-based inferences)的方法来界定，具体表现为某个变量增加两个标准差对比赛结果的影响(每10场比赛中额外取胜或失利的场数)。研究结果显示，在2014年巴西世界杯小组赛的所有48场比赛中，9个变量(射门、射正、反击中射门、禁区内射门、控球、短传、连续传球平均次数、高空球争抢成功率和抢断)与赢球概率有清晰的正相关关系，4个变量(射门被封堵、传中、过人和红牌)与赢球概率有清晰的负相关关系，其他12个被分析的变量与赢球概率的相关关系微小或不清晰。而在38场小分差比赛中，高空球争抢成功率由正相关变为微小关系，黄牌则由微小关系变为清晰的负相关。在西甲联赛中，每一支球队增加两个标准差的“射正球门”可以给每10场比赛带来3.4场额外胜利(99%置信区间±1.0场)，而所有球队作为一个整体，每增加两个标准差的“射正球门”可以给每10场比赛带来1.7场额外胜利(99%置信区间±1.0场)。其他大多数比赛相关事件与比赛结果的相关关系与“控球”相关联。每一支球队增加两个标准差的“控球”将会给每10场比赛带来1.2场额外失利(99%置信区间±1.0场)，而所有球队作为一个整体，每增加两个标准差的“控球”可以给每10场比赛带来1.7场额外胜利(99%置信区间±1.4场)。与客场比赛相对，主场能给球队带来1.9 /10场额外胜利(99%置信区间±0.9场)。 比赛表现档案和模型中得出的研究结果可以为俱乐部、足球队、教练组、表现分析师和运动员提供详细而直接的参考信息。这些信息可用于训练指导、赛前备战、赛中技战术调整和赛后技战术表现分析，也可运用于足球运动员选材、培养和发展。

GTH-UPM system for search on speech evaluation

This paper describes the GTH-UPM system for the Albayzin 2014 Search on Speech Evaluation. Teh evaluation task consists of searching a list of terms/queries in audio files. The GTH-UPM system we are presenting is based on a LVCSR (Large Vocabulary Continuous Speech Recognition) system. We have used MAVIR corpus and the Spanish partition of the EPPS (European Parliament Plenary Sessions) database for training both acoustic and language models. The main effort has been focused on lexicon preparation and text selection for the language model construction. The system makes use of different lexicon and language models depending on the task that is performed. For the best configuration of the system on the development set, we have obtained a FOM of 75.27 for the deyword spotting task.

Ontology Evaluation: a pitfall-based approach to ontology diagnosis

La evaluación de ontologías, incluyendo diagnóstico y reparación de las mismas, es una compleja actividad que debe llevarse a cabo en cualquier proyecto de desarrollo ontológico para comprobar la calidad técnica de las ontologías. Sin embargo, existe una gran brecha entre los enfoques metodológicos sobre la evaluación de ontologías y las herramientas que le dan soporte. En particular, no existen enfoques que proporcionen guías concretas sobre cómo diagnosticar y, en consecuencia, reparar ontologías. Esta tesis pretende avanzar en el área de la evaluación de ontologías, concretamente en la actividad de diagnóstico. Los principales objetivos de esta tesis son (a) ayudar a los desarrolladores en el diagnóstico de ontologías para encontrar errores comunes y (b) facilitar dicho diagnóstico reduciendo el esfuerzo empleado proporcionando el soporte tecnológico adecuado. Esta tesis presenta las siguientes contribuciones: • Catálogo de 41 errores comunes que los ingenieros ontológicos pueden cometer durante el desarrollo de ontologías. • Modelo de calidad para el diagnóstico de ontologías alineando el catálogo de errores comunes con modelos de calidad existentes. • Diseño e implementación de 48 métodos para detectar 33 de los 41 errores comunes en el catálogo. • Soporte tecnológico OOPS!, que permite el diagnstico de ontologías de forma (semi)automática. De acuerdo con los comentarios recibidos y los resultados de los test de satisfacción realizados, se puede afirmar que el enfoque desarrollado y presentado en esta tesis ayuda de forma efectiva a los usuarios a mejorar la calidad de sus ontologías. OOPS! ha sido ampliamente aceptado por un gran número de usuarios de formal global y ha sido utilizado alrededor de 3000 veces desde 60 países diferentes. OOPS! se ha integrado en software desarrollado por terceros y ha sido instalado en empresas para ser utilizado tanto durante el desarrollo de ontologías como en actividades de formación. Abstract Ontology evaluation, which includes ontology diagnosis and repair, is a complex activity that should be carried out in every ontology development project, because it checks for the technical quality of the ontology. However, there is an important gap between the methodological work about ontology evaluation and the tools that support such an activity. More precisely, not many approaches provide clear guidance about how to diagnose ontologies and how to repair them accordingly. This thesis aims to advance the current state of the art of ontology evaluation, specifically in the ontology diagnosis activity. The main goals of this thesis are (a) to help ontology engineers to diagnose their ontologies in order to find common pitfalls and (b) to lessen the effort required from them by providing the suitable technological support. This thesis presents the following main contributions: • A catalogue that describes 41 pitfalls that ontology developers might include in their ontologies. • A quality model for ontology diagnose that aligns the pitfall catalogue to existing quality models for semantic technologies. • The design and implementation of 48 methods for detecting 33 out of the 41 pitfalls defined in the catalogue. • A system called OOPS! (OntOlogy Pitfall Scanner!) that allows ontology engineers to (semi)automatically diagnose their ontologies. According to the feedback gathered and satisfaction tests carried out, the approach developed and presented in this thesis effectively helps users to increase the quality of their ontologies. At the time of writing this thesis, OOPS! has been broadly accepted by a high number of users worldwide and has been used around 3000 times from 60 different countries. OOPS! is integrated with third-party software and is locally installed in private enterprises being used both for ontology development activities and training courses.