Establishing the excitation spectrum of the nucleon would be a key advance to further our understanding of nucleon structure and Quantum Chromodynamics (QCD). Recent theoretical advances allow predictions of the excitation spectrum of the nucleon and other nucleon properties directly from QCD in the non-perturbative regime, via numerical methods (such as Lattice QCD), complementing existing constituent quark models. There is an ongoing world programme in meson photoproduction from the nucleon, which has already led to a number of nucleon resonances being discovered and established. This advance has largely been made possible by the first accurate measurement of polarisation observables. Available data has been obtained for proton targets, whereas for a complete picture of meson photoproduction, data from the neutron must also be obtained. This is important, as nucleon resonances can have very different photo-couplings to the proton and neutron. This thesis presents the first measurement of the E double-polarisation observable for the exclusive γn → K+Ʃ- reaction using a polarised hydrogen-deuterium target from the g14 run period at CLAS. Circularly polarised photons of energies between 1:1 and 2:3 GeV were used, with results shown in 200 MeV bins in Eγ and bins of 0:4 in cos θC.M./K+. Further to this, CLAS has undergone a detector upgrade in order to facilitate electrons of up to 12 GeV from Jefferson Lab's upgraded accelerator. Essential to this, is a new system for tagging quasi-real photons by detecting electrons scattered at very small angles. My work includes significant contributions to the design, realisation and construction of a hodoscope for this forward photon tagging apparatus. Presented in this thesis is a comprehensive overview of my work in developing and constructing the scintillating hodoscope for the CLAS12 Forward Tagger.

Meson spectroscopy aims to study the masses and decay processes of mesons to better understand the mechanics of Quantum Chromo Dynamics in the strong coupling (non peturbative) regime. The Edinburgh group are spokespersons on a major new proposal at the Thomas Jefferson National Laboratory called MesonEx. This experiment will make use of the newly upgraded 12 GeV electron beam and detector systems to produce and measure the properties of mesons of interest for study. Edinburgh also leads the development of a fast timing Hodoscope, part of the new Forward Tagger detector system that will be essential to the success of this experiment. The primary focus of this thesis is the development process of the Foward Tagger Hodoscope, from its inception up to its installation at Jefferson Lab. The second part of the document will discuss the first measurement of the γp → ωππp decay channel, with data from the g11a run of the CLAS (CEBAF Large Acceptance Spectrometer) in Hall B of Jefferson Lab. The analysis presents evidence for resonant contributions to the dataset decaying directly to ωππ, via b1(1235)π or via ωρ. These include a2(1320), π1(1600), ω3(1670) and π2(1670).

In this work we report on two sets of measurements involving the Hall-B photon tagging system of the Tomas Jefferson National Accelerator Facility. The relative energy calibration of the tagging counters was performed by using the PrimEx pair spectrometer and a series of high position resolution micro-strip detectors. The photon energies were determined by forming coincidence between the tagger and the e+ e- pairs for several values of the pair spectrometer magnetic field between 0.36T to 1.9T (total of 180 fields). The second set of measurements, collected in conjunction with the Fall 2004 PrimEx run, investigated inherent limitations on the photon tagging technique. We report for the first time an experimental signature for these effects.

No description available.

Made available in DSpace on 2016-04-28T18:23:32Z (GMT). No. of bitstreams: 1 Agnes dos Santos Scaramuzzi Rodrigues.pdf: 914037 bytes, checksum: 09d927669b18d6b386fffa399538f2cc (MD5) Previous issue date: 2007-07-10 / Coordenação de Aperfeiçoamento de Pessoal de Nível Superior / The aim of this research study was to look at the language of leadership in some areas such as politics, war, religion and business, across a large span of time by means of the analysis of corpora. The specific objectives are: first, to observe the recurrence of use across leaders, and second to identify the conceptual metaphors that underline the speech of the leaders. Our main theoretical support is Corpus Linguistics, which can be defined as an area that is concerned with the collection and the exploration of corpora, or linguistic data sets (BERBER SARDINHA, 2004: 3). In addition to Corpus Linguistics, the study presented here drew on conceptual metaphors (according to Lakoff and Johnson, 1980 / 2002). The central claim is that conceptual metaphors structure the way we conceive the world. The corpora used in the research were formed by samples of the discourse of the following people: Hitler, Getúlio D. Vargas, Napoleon Bonaparte, Saint Paul, five executives, two consultants and a president of a local company. The number of words of the corpus is of 317,757 words (tokens), that is, a medium-sized corpus. The results indicated, in answer to the first research question, that the words found in the corpus with higher metaphorical probability across the majority of the leaders were ten: time ; had ; life ; name ; example ; fact ; to be ; company and work . Of these words two were analyzed in depth: time and life . In relation to the second question, for the word time we found the following metaphors conceptual: TIME IS AN INDETERMINATE AMOUNT; TIME IS INDETERMINATE SPACE and TIME IS A VALUABLE POSSESSION, and for the word life we found: CULTURE IS AN ORGANISM; THE ECONOMY IS AN ORGANISM; POLITICS IS AN ORGANISM; POLITICAL CAREER IS AN ORGANISM; A PROFESSION IS AN ORGANISM; THE NATION IS AN ORGANISM and INSTITUTIONS ARE ORGANISMS. In view of these findings, we concluded, first, that the words time and life are resources for the expression of leadership. The analysis of the data revealed in the analyses of the agreements we observe that 43.96% of the occurrences of time and 51.70% of the occurrences of life were metaphorical. Moreover, our findings indicate similarities between the discourse of the leaders in different spheres of human activity. Thus, we assume that these similarities can be considered a characteristic of the discourse of leadership / O objetivo geral desta pesquisa é estudar a linguagem de liderança em várias áreas de atuação humana: política; militar; religião e negócios, em épocas diferentes pelo viés da metáfora e por meio da análise de corpora. Os objetivos específicos são: observar a recorrência de uso metafórico das palavras pela maioria dos líderes e identificar aquelas com maior probabilidade de usos metafóricos e, a partir dessas palavras, identificar que metáforas conceptuais são realizadas no discurso dos líderes. Adotamos como suporte teórico a Lingüística de Corpus definida como uma área de pesquisa que se ocupa da coleta e da exploração de corpora, ou conjunto de dados lingüísticos textuais coletados criteriosamente, com o propósito de servirem para a pesquisa de uma língua ou variedade lingüística dedicando-se à exploração da linguagem , escrita ou oral, por meio de evidências empíricas extraídas de computador (BERBER SARDINHA, 2004: 3). A Teoria da Metáfora Conceptual de Lakoff e Johnson (1980 / 2002) foi outra área de pesquisa que fundamentou este estudo. Essa teoria propõe que nosso sistema conceptual é regido por meio das metáforas, até mesmo nossas atividades cotidianas são orientadas por elas. Os corpora empregados foram: Subcorpus Histórico e Subcorpus Empresarial. O Subcorpus Histórico foi composto por: Hitler, Getúlio D. Vargas, Napoleão Bonaparte e São Paulo. O Subcorpus Empresarial por: cinco executivos, dois consultores e um presidente de empresa nacional. O número de palavras do corpus é de 317.757 palavras, isto é, um corpus médio. Os resultados indicaram que as palavras encontradas no corpus com maior probabilidade metafórica de usos entre a maioria dos líderes são dez: tempo ; tinha ; vida ; nome ; exemplo fato ; ser ; empresa e trabalho . Dessas palavras, duas foram analisadas detalhadamente, sendo: tempo e vida . Para a palavra tempo , encontramos as seguintes metáforas conceptuais: TEMPO É QUANTIDADE INDETERMINADA; TEMPO É ESPAÇO INDETERMINADO e TEMPO É UM BEM VALIOSO e para a palavra vida encontramos metáforas do organismo, isto é, A CULTURA; A ECONOMIA; A POLÍTICA; A CARREIRA POLÍTICA; A PROFISSÃO; A NAÇÃO, todas entendidas como organismos, por exemplo, A INSTITUIÇÃO É UM ORGANISMO. Concluímos que as palavras tempo e vida são recursos de expressão da liderança. Nas análises das concordâncias, observamos que 43,96% das ocorrências de tempo eram metafóricas e que 51,70% das ocorrências de vida também o eram

Because the development of resources in any language is an expensive process, many languages, including the indigenous languages of South Africa, can be classified as being resource scarce, or lacking in tagging resources. This study investigates and applies techniques and methodologies for optimising the use of available resources and improving the accuracy of a tagger using Afrikaans as resource-scarce language and aims to i) determine whether combination techniques can be effectively applied to improve the accuracy of a tagger for Afrikaans, and ii) determine whether structural semi-supervised learning can be effectively applied to improve the accuracy of a supervised learning tagger for Afrikaans. In order to realise the first aim, existing methodologies for combining classification algorithms are investigated. Four taggers, trained using MBT, SVMlight, MXPOST and TnT respectively, are then combined into a combination tagger using weighted voting. Weights are calculated by means of total precision, tag precision and a combination of precision and recall. Although the combination of taggers does not consistently lead to an error rate reduction with regard to the baseline, it manages to achieve an error rate reduction of up to 18.48% in some cases. In order to realise the second aim, existing semi-supervised learning algorithms, with specific focus on structural semi-supervised learning, are investigated. Structural semi-supervised learning is implemented by means of the SVD-ASO-algorithm, which attempts to extract the shared structure of untagged data using auxiliary problems before training a tagger. The use of untagged data during the training of a tagger leads to an error rate reduction with regard to the baseline of 1.67%. Even though the error rate reduction does not prove to be statistically significant in all cases, the results show that it is possible to improve the accuracy in some cases. / MSc (Computer Science), North-West University, Potchefstroom Campus, 2015

Because the development of resources in any language is an expensive process, many languages, including the indigenous languages of South Africa, can be classified as being resource scarce, or lacking in tagging resources. This study investigates and applies techniques and methodologies for optimising the use of available resources and improving the accuracy of a tagger using Afrikaans as resource-scarce language and aims to i) determine whether combination techniques can be effectively applied to improve the accuracy of a tagger for Afrikaans, and ii) determine whether structural semi-supervised learning can be effectively applied to improve the accuracy of a supervised learning tagger for Afrikaans. In order to realise the first aim, existing methodologies for combining classification algorithms are investigated. Four taggers, trained using MBT, SVMlight, MXPOST and TnT respectively, are then combined into a combination tagger using weighted voting. Weights are calculated by means of total precision, tag precision and a combination of precision and recall. Although the combination of taggers does not consistently lead to an error rate reduction with regard to the baseline, it manages to achieve an error rate reduction of up to 18.48% in some cases. In order to realise the second aim, existing semi-supervised learning algorithms, with specific focus on structural semi-supervised learning, are investigated. Structural semi-supervised learning is implemented by means of the SVD-ASO-algorithm, which attempts to extract the shared structure of untagged data using auxiliary problems before training a tagger. The use of untagged data during the training of a tagger leads to an error rate reduction with regard to the baseline of 1.67%. Even though the error rate reduction does not prove to be statistically significant in all cases, the results show that it is possible to improve the accuracy in some cases. / MSc (Computer Science), North-West University, Potchefstroom Campus, 2015