Name Entity Recognition

Results of queries by personal names often contain documents related to several people because of the namesake problem. In order to differentiate documents related to different people, an effective method is needed to measure document similarities and to find documents related to the same person. Some previous researchers have used the vector space model or have tried to extract common named entities for measuring similarities. We propose a new method that uses Web directories as a knowledge base to find shared contexts in document pairs and uses the measurement of shared contexts to determine similarities between document pairs. Experimental results show that our proposed method outperforms the vector space model method and the named entity recognition method.

Results of queries by personal names often contain documents related to several people because of the namesake problem. In order to differentiate documents related to different people, an effective method is needed to measure document similarities and to find documents related to the same person. Some previous researchers have used the vector space model or have tried to extract common named entities for measuring similarities. We propose a new method that uses Web directories as a knowledge base to find shared contexts in document pairs and uses the measurement of shared contexts to determine similarities between document pairs. Experimental results show that our proposed method outperforms the vector space model method and the named entity recognition method.

recibido: 09 de noviembre de 2013, aceptado: 02 de enero de 2014 RESUMEN. Un Índice de Calidad de Agua (ICA) es una herramienta estadística para estimar la calidad de un cuerpo de agua. El objetivo fue determinar un ICA para la presa La Boquilla en Chihuahua, México. Seis puntos de muestreo fueron seleccionados aleatoriamente y muestras de agua se colectaron mensualmente, a cuatro profundidades: 0.30 m, 5.0 m, 10.0 m y 15.0 m. Se cuanticó el potencial hidrógeno (pH), oxígeno disuelto (OD), conductividad eléctrica (CE), sólidos disueltos totales (SDT), temperatura (T), cloruros (Cl-), dureza total (DT) y turbidez. El análisis estadístico consideró un arreglo de tratamientos factorial 12 × 4 donde el factor A fue el tiempo de muestreo (12 meses) y el factor B la profundidad (cuatro profundidades). El ICA fue calculado para 0.30 m y 15.0 m de profundidad aplicando una ecuación universal. Los rangos de las variables evaluadas fueron como sigue:

Los avances tecnológicos de los sistemas de cómputo paralelo y distribuido permiten el desarrollo de aplicaciones antes impensadas. Una de nuestras líneas de investigación se enfoca en aplicar estas tecnologías a Unidades de Cuidados Intensivos y Unidades de Vigilancia Intermedia. Buscamos mejorar sustancialmente el rendimiento de ellas con el procesamiento en tiempo real de datos masivos generados por el equipamiento médico y otras fuentes. Adicionalmente, trabajamos en la reducción del consumo energético de los sistemas de computación de altas prestaciones, con especial atención en los mecanismos de tolerancia a fallos. Todas nuestras investigaciones se centran en desarrollar metodologías, modelos y soluciones informáticas para colaborar en la resolución de problemas que tengan una alta demanda computacional e impacto social. Los trabajos se desarrollan en colaboración con otras universidades, y un hospital público de Argentina. La formación de recursos humanos en estas líneas está orientada al nivel de grado, maestría y doctoral. Palabras claves: computación de altas prestaciones, eciencia energética, big data, aplicaciones para la salud.

Semantic Web technologies like RDF model, URIs and SPARQL query language, can reduce the complexity of data integration by making use of properly established and described links between sources. However, the diculty to formulate distributed queries has been a challenge to harness the potential of these technologies due to autonomy, distribution and vocabulary of heterogeneous data sources. This scenario demands eective mechanisms for integrating data on Linked Data. Linked Data Mashups allow users to query and integrate structured and linked data on the web. This work proposes an architecture of Linked Data Mashups based on the use of Linked Data Mashup Services (LIDMS). A module for ecient execution of federated query plans on Linked Data has been developed and is a component of the proposed architecture. The results of experiments using the execution module were more ecient than other existing strategies. Furthermore, a LIDMS execution Web environment also has been dened and implemented as contribution of this work. Resumo. Tecnologias da Web Semântica como modelo RDF, URIs e linguagem de consulta SPARQL, podem reduzir a complexidade de integração de dados ao fazer uso de ligações corretamente estabelecidas e descritas entre fontes. No entanto, a diculdade para formulação de consultas distribuídas tem sido um obstáculo para aproveitar o potencial dessas tecnologias em virtude da autonomia, distribuição e vocabulário heterogêneo das fontes de dados. Esse cenário demanda mecanismos ecientes para integração de dados sobre Linked Data. Linked Data Mashups permitem aos usuários executar consultas e integrar dados estruturados e vinculados na web. O presente trabalho propõe uma arquitetura de Linked Data Mashups baseada no uso de Linked Data Mashup Services (LIDMS). Um módulo para execução eciente de planos de consulta federados sobre Linked Data foi desenvolvido e é um componente da arquitetura proposta. Os resultados de experimentos realizados com o uso do módulo de execução mostraram-se mais ecientes que outras estratégias existentes. Além disso, um ambiente Web para execução de LIDMS também foi denido e implementado como contribuição deste trabalho.

Semantic Web technologies like RDF model, URIs and SPARQL query language, can reduce the complexity of data integration by making use of properly established and described links between sources. However, the diculty to formulate distributed queries has been a challenge to harness the potential of these technologies due to autonomy, distribution and vocabulary of heterogeneous data sources. This scenario demands eective mechanisms for integrating data on Linked Data. Linked Data Mashups allow users to query and integrate structured and linked data on the web. This work proposes an architecture of Linked Data Mashups based on the use of Linked Data Mashup Services (LIDMS). A module for ecient execution of federated query plans on Linked Data has been developed and is a component of the proposed architecture. The results of experiments using the execution module were more ecient than other existing strategies. Furthermore, a LIDMS execution Web environment also has been dened and implemented as contribution of this work. Resumo. Tecnologias da Web Semântica como modelo RDF, URIs e linguagem de consulta SPARQL, podem reduzir a complexidade de integração de dados ao fazer uso de ligações corretamente estabelecidas e descritas entre fontes. No entanto, a diculdade para formulação de consultas distribuídas tem sido um obstáculo para aproveitar o potencial dessas tecnologias em virtude da autonomia, distribuição e vocabulário heterogêneo das fontes de dados. Esse cenário demanda mecanismos ecientes para integração de dados sobre Linked Data. Linked Data Mashups permitem aos usuários executar consultas e integrar dados estruturados e vinculados na web. O presente trabalho propõe uma arquitetura de Linked Data Mashups baseada no uso de Linked Data Mashup Services (LIDMS). Um módulo para execução eciente de planos de consulta federados sobre Linked Data foi desenvolvido e é um componente da arquitetura proposta. Os resultados de experimentos realizados com o uso do módulo de execução mostraram-se mais ecientes que outras estratégias existentes. Além disso, um ambiente Web para execução de LIDMS também foi denido e implementado como contribuição deste trabalho.

dos (Knowledge Discovery in Databases -KDD) envolve a análise de extensas bases de dados e recurso a complexos algoritmos de análise de dados (Data Mining ). Este processo requer, geralmente, recursos computacionais dedicados e de elevado custo o que reduz signicativamente o número de utilizadores capazes de efectuar tais análises. Neste artigo apresentamos uma arquitectura baseada em computadores pessoais distribuídos numa rede de computadores de uma organização e que permite a realização de tarefas de KDD sem recursos computacionais dedicados e sem perturbar o funcionamento da organização. A arquitectura é denominada Harvard -HARV esting Architecture of idle machines foR Data mining. O Harvard utiliza uma linguagem de especicação e controlo de tarefas baseada em XML. A linguagem XML no caso do Harvard é imprescindível para a interoperabilidade entre os diferentes componentes do ambiente descrevendo claramente todos os aspectos da tarefa de KDD a ser executada de forma distribuída. Os resultados alcançados por diferentes nós do sistema são transcritos em XML, de modo a facilitar a apresentação ao utilizador do ambiente Harvard e ainda permitir integração com outros sistemas de extracção de conhecimento. O crescente uso das tecnologias da informação associado ao crescimento e desenvolvimento económico de vários sectores (empresarial, governo, comunidade cientíca e académica, entre outros) têm permitido que as organizações e a sociedade em geral reúnam uma enorme quantidade de dados e informações sistematicamente em Bases de Dados (BD) [?]. Do ponto de vista humano, torna-se praticamente impossível interpretar, analisar e obter resultados a partir de grandes quantidades de dados sem o auxílio de computadores. Torna-se muito difícil diferenciar informações verdadeiramente importantes e úteis em tão grande quantidade de dados [?,?]. Neste caso, a aplicação de técnicas automáticas de análise capazes de extraír informação útil torna-se inevitável dada a complexidade e a extensão das BD.

Artículo recibido: 23 de noviembre de 2009, aceptado: 08 de abril de 2013 RESUMEN. Un total de 12 mieles fueron colectadas durante los ciclos de cosecha 2006-2007 en los municipios de Huimanguillo, Cárdenas, Paraíso (región de La Chontalpa), Centla (región de los Pantanos) y Tacotalpa (región de La Sierra), del estado de Tabasco, con el objetivo de caracterizarlas botánicamente. Para la caracterización de las mieles se emplearon técnicas melisopalinológicas (análisis del polen) y técnicas sicoquímicas (pH, cenizas y conductividad eléctrica). Los análisis sicoquímicos no mostraron diferencias entre las muestras de miel, por lo que no pudieron ser útiles para su clasicación, mientras que la caracterización palinológica permitió identicartres mieles monoorales de Cocos nucifera (Centla), Mimosa orthocarpa var. berlandieri (Paraiso) y Psidium sp. (Tacotalpa), siete mieles poliorales y dos mieles biorales de Acalypha sp. / Bursera simaruba (Tacotalpa) y Gramineae / Celtis sp. (Centla). Estos resultados muestran que en un mismo apiario se pueden cosechar dos diferentes tipos de miel según la temporada de oración. En general, en Tabasco se producen diferentes tipos de miel dependiendo de la región geográca, aún cuando dicho estado no destaca en volumen de producción de miel, se comprobó que si podría competir a nivel comercial en la producción de mieles monoorales que son apreciadas principalmente en la Unión Europea. Palabras clave: caracterización de miel, Apis mellifera, Tabasco, melisopalinología, monooral, bioral y multioral. ABSTRACT.A total of 12 honey samples were collected for botanical characterisation during the 2006-2007 harvest cycle from the municipalities of Huimanguillo, Cárdenas, Paraíso (Chontalpa region), Centla (Wetlands region) and Tacotalpa (Sierra region) in the state of Tabasco. Melisopalynological (pollen analysis) and physicochemical (pH, ash and electrical conductivity) techniques were used to characterise the honey samples. The physicochemical analyses showed no dierences among the honey samples, for which reason they were not useful. In contrast, the palynological characterisation made it possible to identify three unioral honey samples of Cocos nucifera (Centla), Mimosa orthocarpa var. berlandieri (Paraíso) and Psidium sp (Tacotalpa), seven multioral honey samples and two bioral honey samples of Acalypha sp / Bursera simaruba (Tacotalpa) and Gramineae / Celtis sp (Centla). The results indicate that, according to the blooming season, it is possible to harvest two dierent types of honey in one apiary. Tabasco in general produces dierent types of honey in dierent geographic regions and, although the state has a low production of honey, it was proved that it may compete commercially in the production of unioral honey that is well received, particularly in the European Union

Fruits from six selections (S) of mamey sapote were harvested in the Soconusco, Chiapas at a mature stage and ripened at room temperature (23 oC and 70 % R. H.) for 12 d. The fruits from the six selections showed maximum respiration and ethylene rates between 31 and 44 mL CO2 kg −1 h−1, and 225 and 550 μL kg−1 h−1, respectively. Fruits from selections S1 and S2 kept esh rmness for longer (≤ 20 N), lowest content of phenolic compounds (0.5-0.6 mg g−1 fresh weight) and the best color parameters (L* = 61, C* = 46 and h = 55). Soluble solids content was between 17.5 y 23 ◦Brix and total sugars was 122 y 196 mg g−1 at ripe stage of selections evaluated with no signi cant di erences. We conclude the S1 and S2 selections had the best postharvest characteristics.

Bu çalışmanın amacı, bir dokümandaki en önemli cümleleri seçerek ilgili dokümanın özetini çıkarmaktır. Bu amaçla 15 farklı cümle seçim metodu kullanılmıştır. Bu metotlar, 15 kadın ve 15 erkek olmak üzere, toplam 30 kişi tarafından çıkarılmış özet dokümanlarının oluşturduğu bir değerlendirme veri seti üzerinde kıyaslanmıştır. Ayrıca, bu metotlardan en başarılı olanlarının birlikte kullanılması ile elde edilen farklı özellik gruplarının başarım değerleri sergilenmiş ve analiz sonuçları paylaşılmıştır. Anahtar Kelimeler-metin özetleme; cümle seçim metotları; varlık ismi tanıma.

Good Computing Reports (From Charles Hu , "Practical Guidance for Teaching the Social Impact Statement (SIS). From Proceedings of the 1996 Symposium on Computers and the Quality of Life, pp. 86-89. New York, ACM Press.) Key Links 1. Materials from Magic Copy Center: Good Computing: A Virtue Ethics Approach to Computer Ethics, Chapter Two, Hu /Frey 2. www.computingcases.org Goals: 1. To uncover ethical surprises in major design projects. (These are ethical issues potential ethical problems that are embedded in the design project.) 2. To communicate e ectively to the client the importance of considering ethical issues and problems associated with design projects upstream in the design process. (This means raising ethical problems from the beginning of the design process and continuously throughout the design process. This is opposed to the idea of waiting until the design process is nished to raise ethical issues.) Four Presuppositions 1. Socio-technical systems and their compone...

ADD elgeri heision higrmF CTL gomputtion ree vogiF CTMC gontinuousEime wrkov ghinF DTMC hisreteEime wrkov ghinF JML tv wodeling vngugeF MDP wrkov heision roessF PCTL roilisti gomputtion ree vogiF PMC rmetri wrkov ghinF PVS rototype eri(tion ystemF SPL oftwre rodut vineF TCC ypeEorretness gonditionF UML ni(ed wodeling vngugeF xxi Contents List of Figures xiii List of Tables xvi List of Denitions xvii List of Theorems xix Acronyms xxi

Resumo | O turismo fornece às pessoas idosas uma atividade social que melhora a qualidade de vida. Emergiu, assim, um produto que tem ganho grande visibilidade: o turismo de saúde. Apesar da importância, ainda não é possível determinar o seu impacto em Portugal e para o mercado sénior. O estudo tem como objetivos retratar a situação atual, em Portugal, e na região Norte, das políticas públicas dirigidas ao turismo de saúde para pessoas idosas e criar as bases para uma política promotora direcionada para seniores em Portugal. Optou-se por realizar um estudo qualitativo exploratório, com recurso a entrevistas semiestruturadas. A amostra foi constituída por 10 atores-chave no âmbito. Os resultados indicam que: (i) existe um parco conhecimento dos decisores sobre o conceito de turismo de saúde e as suas divisões; (ii) Portugal não tem denida uma estratégia especíca para o setor; (iii) o envelhecimento é visto como uma oportunidade para o fenómeno; (iv) as intervenções que devem ser realizadas são o desenvolvimento de uma estratégia concertada, a promoção internacional e a identicação dos intervenientes na área do turismo de saúde. Os resultados obtidos permitem conhecer melhor a situação atual no contexto do turismo de saúde em Portugal (o que falta fazer, o que se está a fazer, como e por quem), assim como o conhecimento que os decisores da área têm no que diz respeito a conceitos e boas-práticas afetas ao tema. Palavra-chave | Envelhecimento, turismo sénior, turismo de saúde, políticas públicas Abstract | Tourism provides to the elderly a social activity that improves their quality of life. Emerged, in that context, a product that has gained great visibility: health tourism. Despite the importance it is still not possible to determine its impact in Portugal, and for the senior market. The objective of the study was to portray the current situation in Portugal and in the North of the public policies aimed at health tourism for the elderly, and to create bases for a promotional policy aimed at seniors in Portugal. It was decided to conduct a qualitative exploratory study, using semi-structured interviews. The sample

Resumo | Este trabalho tem como objetivo analisar o conceito de turismo académico, inserido na esfera do turismo educacional, e aplicar o conceito aos estudantes que realizaram mobilidade internacional ao abrigo do Programa Erasmus na Universidade de Aveiro, entre 2006 e 2012. Para tal, foi aplicado um inquérito por questionário a estes estudantes que permitiu concluir que a sua experiência positiva é na-turalmente enriquecedora para a Universidade de Aveiro, permitindo a captação crescente de estudantes internacionais, tendo consequentemente uma importância económica e cultural de grande impacte na cidade, por extravasamento natural do seu campus. Palavras-chave | Turismo académico, turismo jovem, estudantes Erasmus, Universidade de Aveiro Abstract | The main objective of this work is to analyze the concept of academic tourism, within the sphere of educational tourism, and apply this concept to international mobility exchange students who undertook the Erasmus Program at the Univer...

Ao professor Amilton Sinatora pela oportunidade para vir ao Brasil e ter uma inesquecível experiência além do doutorado, pela amizade, apoio, conança e motivação para realizar este trabalho. À FAPESP pela bolsa de doutorado (processo 2005/59131-0) que permitiu o desenvolvimento deste projeto, a participação nos congressos e a minha manutenção. Aos professores César Chaves, Juan Manuel Vélez e Alejandro Toro que contribuíram bastante para seguir este caminho. À minha amiga Diana López pelo recebimento no Brasil, que fez mais suportável deixar longe a namorada, a família e os amigos. À minha sogra Candida Castillo e às famílias Arango Gómez e Viáfara Quintana pelo carinho e ajuda. Aos amigos de Medellín e outros lugares pela maravilhosa amizade. Aos amigos e doutores Mario Vitor Leite, Abel Recco, Vanderlei Ferreira e María Cristina Moré Farias pela amizade e experiência compartilhada.

Phosphate glasses having compositions (59.5x)P2O540MgOxAgCl0.5Er2O3, where x = 0, 1.5 mol.% is prepared using melt-quenching technique. Infrared, absorption and photoluminescence spectra of Er 3+-doped magnesium phosphate glasses have been reported. The amorphous nature of the host glass is conrmed by X-ray diraction technique. Transmission electron microscope image conrms the existence of silver nanoparticles inside the glass matrix. The localized surface plasmon resonance band of silver is found to be located around ≈ 528 nm for the Er 3+ free sample. A frequency upconversion process from infrared to visible is observed on excitation with 797 nm radiation. Furthermore, an enhancement in the emission at 540 nm and 632 nm is found due to the local eld eect of silver nanoparticles. Our ndings may contribute towards the development of solid state laser and sensors.

En el contexto de la propagación del coronavirus (Covid-19), a nivel mundial, se han reportado datos sobre el número de casos positivos, fallecidos, hospitalizados, recuperados, etc. En el caso de Perú por el Ministerio de Salud (MINSA), a partir de estos se realiza un análisis poniendo énfasis en el número de casos positivos y fallecidos y tomando como referencia el comportamiento de un fenómeno físico se postula un número f, para describir la evolución de los casos positivos y fallecidos por coronavirus.

Denna rapport fokuserar på jämförelsen av några olika klassificeringsmetoder applicerade på bilddatan Fashion-MNIST. De olika metoderna är artificiella neurala nätverk och funktionell principalkomponentanalys och principalkomponentanalys. För de neurala nätverken har vi två typer: CNN och FNN. Den förstnämnda är specialiserad på just bilder medan den sistnämnda kan appliceras på olika typer av dataset fast har nackdelen med försämrad noggrannhet. Funktionell principalkomponentanalys eller FPCA är en utvidgning av principalkomponentanalys (PCA) som innebär studiet av dimensionreducering av högdimensionell data. FPCA uttrycker data i form av funktioner vilket möjliggör ytterligare dimensionsreducering om funktionerna effektivt representerar datan. Parametrar i majoriteten av metoderna bestäms med hjälp av korsvalidering. Korsvalidering tillämpas för att en modell inte ska bli partiskt mot en viss del av datan och vi kan på så vis kan dra rättvisa slutsatser. Resultaten ger att de neur...

The paper deals with one small step in the process of model driven development (MDD) or model driven architecture (MDA) widely used terms nowadays. MDD denes techniques to develop software systems using variety of models together with a set of transformations. MDD species several levels of models depending on abstraction ranging from computation independent models (CIM) to platform independent models (PIM) into platform specic models (PSM), and implementation specic models (ISM). Many CASE tools provide automated support for generating more specic models from more abstract ones e.g. PSM or ISM from PIM. This task is referred to as forward engineering. Many tools also provide automated support for generating more abstract models from more specic ones e.g. PIM from PSM or ISM. This task is referred to as reverse engineering. Some tools also support round-trip engineering, which involves both forward and reverse engineering steps such that software artifacts (model and code) become synchronized. However, these tools need exact transformation rules to be dened for such transformations. This paper describes basic principles and restrictions for transformations of binary relationships and transformations of binary relationships with the particular multiplicity from PIM level into PSM level. The idea is illustrated on examples.

Track etched membranes are porous systems consisting of a thin polymer foil with channels from surface to surface. Latent ion tracks are the result of the passage of swift ions through solid matter and they can be etched selectively. As a result, conical, cylindrical or other shape channels can be obtained. The increasing interest in the polymer track etched membranes with nanochannels is connected with development and creation of nanoporous materials of unique properties. The template synthesis method based on deposition of materials inside well-dened uniform pores of membranes oers unique possibilities of formation of one-dimensional, high aspect ratio (length to diameter) cylindrical species having form of rods, wires, tubules, multiwall tubules and multilayer rods, practically from any solid material. Metal-organic frameworks are a class of hybrid materials comprising metal ion-based vertices and organic ligands (linkers) which serve to connect the vertices into one-, two-or three-dimensional periodic structures. A specic property of porous structures is their intrinsic porosity, which renders them potentially useful for gas storage, separation and catalysis. The possibility of obtaining a new composite material: polymeric track etched membrane with pores lled with hybrid porous material has been demonstrated.

Ocimum basilicum belongs to the family Lamiaceae which is known to have anticancer and many other bioactivities. Phoma eupyrena, Emericella nidulans lata and Chaetomium olivaccium were isolated from the different organs of the Basil plant. Root, stem and leaf’s endophytes extracts of this plant, were tested for phytochemical constituents, antioxidant and anti- α-amylase activities. Phytochemical screening of fungal endophyte extracts revealed the presence of alkaloids, steroids, terpenoids and saponins.  Quantitative determination of total phenolics, total flavonoids and various in vitro antioxidant  activities (DPPH, H2O2, α-Amylase,  reducing power and antioxidant capacity) of methanolic extract was carried out using colorimetric methods.  The results of this study displayed that antioxidant activities differ significantly in the endophyte’s extract of the plant. The flavonoid content appeared in phoma extract only. The  IC50 values  for  DPPH  radical scavenging was in the order ...

A ship design with boundary layer alignment device (BLAD) aiming at improving the inow of a propeller is evaluated using a RANSE computation with adaptive grid renement. This paper is focused on model scale simulation for which experimental results can be used for validation. Propeller performance is evaluated with a RANSE/BEM coupling procedure. Numerical simulation reveals that adaptive grid renement is an ecient way to obtain a reliable prediction for such conguration. Moreover, grid renement criterion is found to be of crucial importance. The newly proposed smoothed pressure Hessian is capable of evaluating correctly the impact of the BLAD on the performance of the propeller.

From 18.04.04 to 23.04.04, the Dagstuhl Seminar 04171 Logic Based Information Agents was held in the International Conference and Research Center (IBFI), Schloss Dagstuhl. During the seminar, several participants presented their current research, and ongoing work and open problems were discussed. Abstracts of the presentations given during the seminar as well as abstracts of seminar results and ideas are put together in this paper. The rst section describes the seminar topics and goals in general. Links to extended abstracts or full papers are provided, if available. 04171 Summary Logic Based Information Agents Logic-based methods have a great potential as a toolbox for the development of information technology infrastructure which is able to provide and handle advances services. With the creation of the world wide web, and the advent of the internet as a communication backbone for connecting people, more and more data and information is becoming available; however, our current meth...

Agradeço a Deus, primeiramente, por conseguir concluir esta tese sem qualquer prejuízo ao Lucas. Agradeço muito ao Diego pelo companheirismo e pela grande compreensão demonstrada em relação a todos os momentos que estive ausente em decorrência deste trabalho. Seu incentivo para conclusão da tese foi muito importante. Meu agradecimento especial à minha mãe, Adélia, sem a qual eu não conseguiria levar adiante este projeto prossional. Enquanto eu trabalhava, ela cuidava com carinho sem igual da minha família. Também agradeço à tia Marli pela atenção que recebi todas as vezes que fui ao Rio de Janeiro por razões relacionadas a este trabalho. Agradeço igualmente à tia Martha, que com frequência vinha nos ajudar a cuidar do Vítor, e, desta forma, permitiu que eu usasse esses momentos para me dedicar à tese. Agradeço à minha orientadora, Prof. Vera, pelas suas preciosas contribuições, disponibilidade e incentivo no decorrer deste trabalho. Agradeço também pela compreensão em relação aos momentos que tive que me afastar para cuidar do Vítor. Agradeço ao Prof. Ruy por gentilmente ter me recebido na PUC-Rio e permitido, na época, meu acesso ao processador de papéis de semânticos F-EXT-WS, o qual utilizo nesta tese. Agradeço também ao Prof. António Branco e a sua equipe de pesquisadores. Todos que me receberam muito bem e com muita atenção, quando estive em Lisboa, por questões relativas a um projeto conjunto entre o nosso grupo de pesquisa e o dele, durante o doutoramento. Agradeço também aos professores da banca por terem aceito a tarefa de avaliar meu trabalho. Agradeço, ainda, ao meu amigo Marcelo Cohen pela sua amizade e inestimável ajuda. Ele sempre se dispôs a me auxiliar nos momentos em que eu precisava resolver problemas relativos à conversão de algum arquivo, conguração de ferramentas e dúvidas relacionadas ao Lyx e ao L A T E X. Agradeço também ao Marco Gonzalez por ter disponibilizado a ferramenta FORMA e resolvido dúvidas quanto ao seu uso. Agradeço ao Marco Mangan por ter me indicado um artigo que foi fundamental para a organização dos capítulos referentes à tese. Agradeço aos meus amigos e colegas pelos momentos de descontração na sala dos professores horistas da PUCRS, os quais tornaram o caminho até a conclusão da tese mais ameno. São eles:

A Deus, que sempre esteve com a gente em cada viagem e em cada momento de nossa vida. Muito obrigada Deus por dar fortaleza especialmente nos momentos de solidão. Resumo MAMANI, E. Z. S. Cálculo de reputação em redes sociais a partir da colaboração entre os participantes. 2013. 93 f.

La tecnologia movil ha surgido de la necesidad de las personas de llevar consigo un medio de comunicacion con opciones de entretenimiento, una biblioteca y acceso a Internet. Actualmente, se ha masificado el uso de dispositivos moviles inteligentes (smart) que se utilizan como algo mas que un mero medio de comunicacion. Desde un dispositivo movil smart se puede, por ejemplo, tomar fotografias, gestionar cuentas bancarias, localizar restaurantes y comercios cercanos a la ubicacion geografica del usuario, registrar las actividades fisicas diarias, e incluso mostrar indicadores biometricos y de salud. El escenario de verificacion de las Apps es muy distinto y mas exigente que el de las aplicaciones que se ejecutan en PC. En primer lugar, el funcionamiento de las Apps puede variar entre dispositivos, esto significa que las Apps se deben analizar fisicamente en varios modelos de dispositivos. En segundo lugar, cualquier programa de analisis puede corromper el funcionamiento normal de un ...

Nuestra investigación está centrada en dos líneas. Por un lado, el estudio del consumo energético de los sistemas de Cómputo de Altas Prestaciones (HPC, de High Performance Computing) cuya alta demanda energética tiene serias consecuencias nancieras, medioambientales, y en muchos casos también sociales. El aumento de la e ciencia energética de los sistemas de HPC no solo proviene de las nuevas arquitecturas hardware, también está involucrado el software, quien debe gestionar y con gurar el sistema para mantener un determinado equilibrio entre tiempo de ejecución, e ciencia energética y productividad. Es en este aspecto, el software, en el que estamos trabajando. Por otro lado, nos interesa el estudio de aplicaciones que utilizan sistemas de cómputo de altas prestaciones orientadas a su uso en centros de salud. En particular, nos enfocamos en sistemas de monitoreo inteligente para Unidades de Vigilancia Intermedia y Unidades de Terapia Intensiva.

From 05.05. to 08.05.2009, the Dagstuhl Seminar 09192 ``From Quality of Service to Quality of Experience'' was held in Schloss Dagstuhl~--~Leibniz Center for Informatics. During the seminar, several participants presented their current research, and ongoing work and open problems were discussed. Abstracts of the presentations given during the seminar as well as abstracts of seminar results and ideas are put together in this paper. The first section describes the seminar topics and goals in general. Links to extended abstracts or full papers are provided, if available.

Despite the fact that stemming greatly improves Arabic information retrieval performance, yet no standard stemmer emerges in the field of Arabic IR due to some limitations and shortcomings. Among the recurring problems is that the stemmer can reduce unrelated words to the same stem as well as fall short to reduce related words to a common stem. Many studies have suggested Arabic algorithms to address the problem associated with stemming. This paper aims to review the state of the retrieval performance of Arabic Light stemmers based on the main objectives achieved, causes for retrieval success and failure, retrieval measure, the affixes, and methodologies. The results showed that light 10 has better retrieval performance compared to other reviewed Arabic light stemmers.

Named entity recognition (NER) plays a significant role in many applications such as information extraction, information retrieval, question answering, and even machine translation. Most of the work on NER using deep learning was done for non-Arabic languages like English and French, and only few studies focused on Arabic. This paper proposes a semi-supervised learning approach to train a BERT-based NER model using labeled and semi-labeled datasets. We compared our approach against various baselines, and state-of-the-art Arabic NER tools on three datasets: AQMAR, NEWS, and TWEETS. We report a significant improvement in F-measure for the AQMAR and the NEWS datasets, which are written in Modern Standard Arabic (MSA), and competitive results for the TWEETS dataset, which contains tweets that are mostly in the Egyptian dialect and contain many mistakes or misspellings.

Despite the fact that stemming greatly improves Arabic information retrieval performance, yet no standard stemmer emerges in the field of Arabic IR due to some limitations and shortcomings. Among the recurring problems is that the stemmer can reduce unrelated words to the same stem as well as fall short to reduce related words to a common stem. Many studies have suggested Arabic algorithms to address the problem associated with stemming. This paper aims to review the state of the retrieval performance of Arabic Light stemmers based on the main objectives achieved, causes for retrieval success and failure, retrieval measure, the affixes, and methodologies. The results showed that light 10 has better retrieval performance compared to other reviewed Arabic light stemmers.

Name Entity Recognition (NER) is a process of information extraction that seeks to locate atomic elements in text and classify them into predefined categories such as the name of persons, organizations, locations, expressions of times, quantities, monetary values, percentages, etc. In this paper, we present an Odia Named Entity Recognizer which will be very useful in searching of information about not only tourism but also all general domains. Here, we follow the method of conditional random field. This is a machine learning technique based on linguistic rules of Name Entities (NE). It handles nested tagging of name entities with a hierarchical tag set containing forty four attributes (level one), thirty eight attributes (level two) and thirty five attributes (level three) in tag set. We have experimented building Conditional Random Field (CRF) models by training the noun phrases of the training data and it gives encouraging results.

Resumen: Este artículo presenta un sistema para el reconocimiento de entidades con nombre apoyándonos en dos técnicas clásicas de aprendizaje automático: los modelos de Markov y losárboles de decisión. Se han desarrollado varios sistemas en los que hemos investigado el efecto producido por la inclusión de características que no dependen en exceso del idioma utilizado. Los experimentos se han realizado con el corpus del español distribuido para la tarea de reconocimiento de entidades con nombre del CoNLL 2002.

Rule-based approaches are using human-made rules to extract Named Entities (NEs), it is one of the most famous ways to extract NE as well as Machine Learning. The term Named Entity Recognition (NER) is defined as a task determined to indicate personal names, locations, organizations and many other entities. In Arabic language, Big Data challenges make Arabic NER develops rapidly and extract useful information from texts. The current paper sheds some light on research progress in rule-based via a diagnostic comparison among linguistic resource, entity type, domain, and performance. We also highlight the challenges of the processing Arabic NEs through rule-based systems. It is expected that good performance of NER will be effective to other modern fields like semantic web searching, question answering, machine translation, information retrieval, and abstracting systems.

Arabic Natural Language Processing (ANLP) has known an important development during the last decade. Nowadays, several ANLP tools are already developed such as morphological analyzers. These analyzers are often used in more advanced applications such as syntactic parsers, search engines, machine translation systems, etc. However, the choice of a morphological analyzer to use, among others, can be difficult for researchers if they ignore its metrics. In this article, we present the challenges of the benchmark of Arabic morphological analyzers. We present also our solution developed in Java, which allows the benchmark by returning the most common metrics, namely the accuracy, precision, f-measure and execution time. This solution has the advantage of being cross-platform, flexible and allows to be extended to cover new morphological analyzers to compare.

Background: Adverse Drug reactions (ADR) cause a high number of deaths among hospitalized patients in developed countries. Major drug agencies have devoted a great interest in the early detection of ADRs due to their high incidence and increasing health care costs. Reporting systems are available in order for both healthcare professionals and patients to alert about possible ADRs. However, several studies have shown that these adverse events are underestimated. Our hypothesis is that health social networks could be a significant information source for the early detection of ADRs as well as of new drug indications. Methods: In this work we present a system for detecting drug effects (which include both adverse drug reactions as well as drug indications) from user posts extracted from a Spanish health forum. Texts were processed using MeaningCloud, a multilingual text analysis engine, to identify drugs and effects. In addition, we developed the first Spanish database storing drugs as well as their effects automatically built from drug package inserts gathered from online websites. We then applied a distant-supervision method using the database on a collection of 84,000 messages in order to extract the relations between drugs and their effects. To classify the relation instances, we used a kernel method based only on shallow linguistic information of the sentences. Results: Regarding Relation Extraction of drugs and their effects, the distant supervision approach achieved a recall of 0.59 and a precision of 0.48. Conclusions: The task of extracting relations between drugs and their effects from social media is a complex challenge due to the characteristics of social media texts. These texts, typically posts or tweets, usually contain many grammatical errors and spelling mistakes. Moreover, patients use lay terminology to refer to diseases, symptoms and indications that is not usually included in lexical resources in languages other than English.

Bu çalışmanın amacı, bir dokümandaki en önemli cümleleri seçerek ilgili dokümanın özetini çıkarmaktır. Bu amaçla 15 farklı cümle seçim metodu kullanılmıştır. Bu metotlar, 15 kadın ve 15 erkek olmak üzere, toplam 30 kişi tarafından çıkarılmış özet dokümanlarının oluşturduğu bir değerlendirme veri seti üzerinde kıyaslanmıştır. Ayrıca, bu metotlardan en başarılı olanlarının birlikte kullanılması ile elde edilen farklı özellik gruplarının başarım değerleri sergilenmiş ve analiz sonuçları paylaşılmıştır. Anahtar Kelimeler-metin özetleme; cümle seçim metotları; varlık ismi tanıma.

The Morphological analysis tool recognizes and investigates the structure of the words given internally and provides the syntactical and morphological info related to words given as input. The Kannada language is morphologically much richer, inflectional and agglutinative dialect. Constructing the better placed morphological analysis tool for Kannada language is much challenging work. In this work, the Natural Language Processing is discussed to investigate the nouns and verbal words of the language. The Morphological analysis tool is for Kannada verbs and nouns is developed using machine learning approaches and presented the results using various types of eight input data sets. The analyser discussed is constructed making use of regulation-dependent, stripping of suffix based and paradigm-dependent methodologies.

In the era of big data, medical researchers attempt to utilize some analysis techniques like machine learning and text mining on their large-scale corpora to save valuable labor work and time. Consequently, many data analysis platforms are built to support medical professionals such as Pubtator, GeneWays, BioContext, etc. These platforms are helpful to medical entities recognition and relation extraction, but there is not an integrated platform to support researchers' various needs, and medical projects are isolated from each other, which is hard to be shared and reused. As a result, we present an integrated system containing 'name entity recognition', 'document categorization' and 'association extraction'. Besides, we add the concept of 'socialization' making projects reusable for further analyses. A case study of chronic kidney disease was adopted to indicate the effectiveness of the proposed system.

Several tools and resources have been developed to deal with Arabic NLP. However, a homogenous and flexible Arabic environment that gathers these components is rarely available. In this perspective, we introduce SAFAR which is a monolingual framework developed in accordance with software engineering requirements and dedicated to Arabic language, especially, the modern standard Arabic and Moroccan dialect. After one decade of integration and development, SAFAR possesses today more than 50 tools and resources that can be exploited either using its API or using its web interface.

Arabic is the 6th most important language in the world with more than 300 million speakers. Arabic Question Answering systems are gaining great importance due to the increasing amounts of Arabic content on the Internet and the increasing demand for information that regular Information Retrieval techniques cannot satisfy. In spite of the importance of Arabic Question Answering, there is no survey that covers it. This was the motivation to write this paper. In this survey, different Arabic Question Answering systems are demonstrated and the main Question Answering tasks like Question Analysis, Passage Retrieval, and Answer Extraction are explored. The main challenges of Modern Standard Arabic are also explained and how these challenges are approached and classified. Different QA Tools are investigated and Future trends are also highlighted to guide new research in this area. This survey provides guidance for new research in Arabic Question Answering to get up-to-date knowledge about the state of the art approaches in this area, the tools used and created by researchers to build an Arabic QA system, and the future trends and new directions to focus on.

Named entity recognition (NER) plays a significant role in many applications such as information extraction, information retrieval, question answering, and even machine translation. Most of the work on NER using deep learning was done for non-Arabic languages like English and French, and only few studies focused on Arabic. This paper proposes a semi-supervised learning approach to train a BERT-based NER model using labeled and semi-labeled datasets. We compared our approach against various baselines, and state-of-the-art Arabic NER tools on three datasets: AQMAR, NEWS, and TWEETS. We report a significant improvement in F-measure for the AQMAR and the NEWS datasets, which are written in Modern Standard Arabic (MSA), and competitive results for the TWEETS dataset, which contains tweets that are mostly in the Egyptian dialect and contain many mistakes or misspellings.

An information retrieval (IR) system is the core of many applications, including digital library management systems (DLMS). The IR-based DLMS depends on either the title with keywords or content as symbolic strings. In contrast, it ignores the meaning of the content or what it indicates. Many researchers tried to improve IR systems either using the named entity recognition (NER) technique or the words’ meaning (word sense) and implemented the improvements with a specific language. However, they did not test the IR system using NER and word sense disambiguation together to study the behavior of this system in the presence of these techniques. This paper aims to improve the information retrieval system used by the DLMS by adding the NER and word sense disambiguation (WSD) together for the English and Arabic languages. For NER, a voting technique was used among three completely different classifiers: rules-based, conditional random field (CRF), and bidirectional LSTM-CNN. For WSD, an e...