Made available in DSpace on 2015-04-14T14:49:42Z (GMT). No. of bitstreams: 1 437178.pdf: 938995 bytes, checksum: 7f4e4a024eb9af218b4ff88670a9ca88 (MD5) Previous issue date: 2011-03-29 / The advances in technology have made the amount of information available in digital format increase rapidly. This increase reflects on the importance of efficient systems to Information Retrieval (IR), getting the right information when it's requested by users. Thesauri can be associated with IR systems, allowing the system to query not only by the key term, but also by related terms, obtaining related documents that were not retrieved. The manual construction, long and costly process that gave rise to the first thesaurus, shall be performed automatically, using different methods and processes available today. With this motivation, this dissertation proposes to study three cases of automatic thesauri construction. One method uses statistical techniques to identify the best related terms. Another method uses syntactic knowledge, being necessary to extract, besides the grammatical categories of each term, the relations that a verb have with its subject or object. The latter method makes use of syntactic knowledge and semantic knowledge of the terms, identifying non apparent relations. For this, this latter method uses an adaptation of the Latent Semantic Analysis technique. We developed three methods for automatic thesaurus construction using documents from the field of data privacy. The results were applied to an IR system, allowing the evaluation by domain experts. In conclusion, we observed that, in certain cases, it's better to apply techniques that do not use semantic knowledge of the terms, obtaining better results with methods that use only the syntactic knowledge of them. / Com o progresso da tecnologia, a quantidade de informa??o dispon?vel em formato digital tem aumentado rapidamente. Esse aumento se reflete na crescente import?ncia de sistemas de Recupera??o de Informa??es (RI) eficientes, obtendo as informa??es corretas quando requisitadas pelos usu?rios. Tesauros podem ser associados a sistemas de RI, permitindo que o sistema realize consultas n?o apenas pelo termo-chave, mas tamb?m por termos relacionados, obtendo documentos relacionados, que antes n?o eram recuperados. A cria??o manual, processo longo e oneroso que dava origem aos primeiros tesauros, passa a ser realizada automaticamente, atrav?s de diferentes m?todos e processos dispon?veis atualmente. Com esta motiva??o, este trabalho prop?e estudar tr?s processos de constru??o autom?tica de tesauros. Um m?todo utiliza t?cnicas estat?sticas para a identifica??o dos melhores termos relacionados. Outro m?todo utiliza conhecimento sint?tico, sendo necess?rio extrair, al?m das categorias gramaticais de cada termo, as rela??es que um verbo tem com seu sujeito ou objeto. O ?ltimo m?todo faz a utiliza??o de conhecimento sint?tico e de conhecimento sem?ntico dos termos, identificando rela??es que n?o s?o aparentes. Para isso, esse ?ltimo m?todo utiliza uma adapta??o da t?cnica de An?lise Sem?ntica Latente. Foram desenvolvidos estes tr?s m?todos de gera??o tesauros a partir de documentos do dom?nio de privacidade de dados. Os resultados foram aplicados a um sistema de RI, permitindo a avalia??o por especialistas do dom?nio. Como conclus?o, observamos que, em determinados casos, ? melhor a aplica??o de t?cnicas que n?o utilizem conhecimento sem?ntico dos termos, obtendo melhores resultados com m?todos que utilizam apenas o conhecimento sint?tico dos mesmos.

Submitted by Jos? Henrique Henrique (jose.neves@ufvjm.edu.br) on 2017-05-04T19:38:16Z No. of bitstreams: 2 license_rdf: 0 bytes, checksum: d41d8cd98f00b204e9800998ecf8427e (MD5) aline_pereira_costa.pdf: 2041406 bytes, checksum: 9995d567446721eda6ae8ff6dbdbe60c (MD5) / Approved for entry into archive by Rodrigo Martins Cruz (rodrigo.cruz@ufvjm.edu.br) on 2017-05-16T16:59:24Z (GMT) No. of bitstreams: 2 license_rdf: 0 bytes, checksum: d41d8cd98f00b204e9800998ecf8427e (MD5) aline_pereira_costa.pdf: 2041406 bytes, checksum: 9995d567446721eda6ae8ff6dbdbe60c (MD5) / Made available in DSpace on 2017-05-16T16:59:24Z (GMT). No. of bitstreams: 2 license_rdf: 0 bytes, checksum: d41d8cd98f00b204e9800998ecf8427e (MD5) aline_pereira_costa.pdf: 2041406 bytes, checksum: 9995d567446721eda6ae8ff6dbdbe60c (MD5) Previous issue date: 2016 / Este projeto surgiu de uma defici?ncia do Departamento de Computa??o da UFVJM (Universidade Federal dos Vales do Jequitinhonha e Mucuri), em armazenar e recuperar seus of?cios. Tal Departamento possui dificuldades na organiza??o e armazenamento eficiente destes of?cios o que inviabiliza o acesso aos documentos e dispende muito tempo na localiza??o e recupera??o da informa??o. Diante disso, foi desenvolvido um Sistema de Indexa??o Autom?tica utilizando-se t?cnicas da biblioteconomia e t?cnicas computacionais que visa automatizar o processo de indexa??o de novos of?cios, otimizar a recupera??o e democratizar o acesso a informa??o. O banco de dados do Sistema foi constru?do baseado em um vocabul?rio controlado: o tesauro. O tesauro ? um tipo de vocabul?rio controlado, mais complexo, que trabalha com ambiguidade dos termos, sinon?mia, rela??es hier?rquicas e associativas e foi elaborado a partir de conceitos selecionados e seus termos relacionados presentes nos of?cios. O escopo inicial do projeto permeia o espa?o de 2011 a 2014, totalizando 239 of?cios. Para a alimenta??o do Sistema com os novos documentos que surgir?o prop?e-se a auto alimenta??o do tesauro que far? a an?lise de relev?ncia de novos termos nos novos of?cios atrav?s de um algoritmo em constru??o. O sistema estar? em ambiente virtual, para que o acesso seja democratizado e o processo de tratamento de novos of?cios seja automatizado. Sendo positiva a implementa??o do projeto, sugere-se que os demais departamentos da UFVJM utilizem o mesmo sistema para organiza??o dos seus documentos, ganhando agilidade nos processos e satisfa??o do usu?rio final na localiza??o do que procura. / Disserta??o (Mestrado Profissional) ? Programa de P?s-Gradua??o em Educa??o, Universidade Federal dos Vales do Jequitinhonha e Mucuri, 2016. / This project arose from an actual deficiency of the UFVJM's Department of Computer Science (Universidade Federal dos Vales do Jequitinhonha e Mucuri) to store and retrieve archives. This institution has been having to put up with problems to store documents efficiently. This difficulty prevents access to documents and imposes a heavy burden to the department's staff, in terms of time to locate and retrieve information. Therefore, we developed an system to index information efficiently. To this end, we have used techniques borrowed from various areas within information science. Such techniques allowed us to automate the process of indexing new archives, optimize their recovery and, as a result, we have been able to democratize the access to information. The proposed database was built based on a controlled vocabulary: the thesaurus. The thesaurus is a controlled type of vocabulary, more complex, working with ambiguity of terms, synonymy, hierarchical relationships and associativity. It was prepared from selected concepts and terms related to these concepts in the archives. The initial scope of the project covers a time period from 2011 to 2014, totaling 239 offices. To feed the system with new documents yet to emerge we propose to self feeding mechanism. This self-feeding process will lead to the relevant analysis of new terms in the new archives through an algorithm, currently under implementation. The system will be in a virtual environment, which ensures not only the democratic access to information, but also the automatic handling of new documents. In case this project obtains positive feedback, we shall suggest that other UFVJM departments use the same system for organizing documents; hence, decreasing their response time, and improving the experience of their end-users.

Submitted by Elba Lopes (elba.lopes@fecap.br) on 2017-08-16T13:55:58Z No. of bitstreams: 2 MARCELO SANTANA DA SILVA.pdf: 579761 bytes, checksum: 31198b1e756da9a787c417c0e54d6c5a (MD5) license_rdf: 0 bytes, checksum: d41d8cd98f00b204e9800998ecf8427e (MD5) / Made available in DSpace on 2017-08-16T13:55:58Z (GMT). No. of bitstreams: 2 MARCELO SANTANA DA SILVA.pdf: 579761 bytes, checksum: 31198b1e756da9a787c417c0e54d6c5a (MD5) license_rdf: 0 bytes, checksum: d41d8cd98f00b204e9800998ecf8427e (MD5) Previous issue date: 2017-02-22 / The objective of this study is to analyze the factors that determine the spread of the public issues of debentures indexed to the Broad Consumer Price Index (IPCA) in Brazil. Emissions indexed to the IPCA were choose because they are instruments usually used to capture resources of longer maturity by the issuing companies. The database had 245 series of issues occurred between January 2010 and December 2015. Regressions were estimated by ordinary least squares and weighted least squares methods, and the results presented by the last method were more robust. The rating-spread ratio was confirmed in all regressions and the results indicate that this variable explains, by itself, 58% of the spread variation. Other the rating, the results indicate that the main factors that determine the spread of the issues are: collateral, issuer experience, maturity, amount, prestige of the coordinating bank, tax benefits and economic scenario. Due to the results achieved, other issues related to market efficiency were approached, such as agency conflicts, information asymmetry and adverse selection. The results show that the collaterized issues remunerated the investors with a higher spread than the unsecured ones, and this premium ranged from 35 to 38 basis points. The results were interpreted in the context of agency theory and resemble those found by John, Lynch and Puri (2003) for the US corporate bonds market. Finally, the favorable economic scenario, as measured by the Emerging Market Bond Index - Brazil (EMBI + BR), showed negativelycorrelated with the spread, and these results were interpreted as effects of the information asymmetry and adverse selection present in the local market for debt issuance. / O objetivo deste estudo ?? analisar os fatores que determinam o spread das emiss??es p??blicas de deb??ntures indexadas ao ??ndice de Pre??os ao Consumidor Amplo (IPCA) no Brasil. Utilizou-se as emiss??es indexadas ao IPCA por serem instrumentos usualmente empregados na capta????o de recursos de maturidade mais longa pelas empresas emissoras. A base de dados contou com 245 s??ries de emiss??es realizadas entre janeiro de 2010 e dezembro de 2015. As regress??es foram estimadas pelos m??todos de m??nimos quadrados ordin??rios e m??nimos quadrados ponderados, e os resultados apresentados por este ??ltimo foram mais robustos. A rela????o rating-spread foi confirmada em todas as regress??es e os resultados indicam que essa vari??vel explica, isoladamente, 58% da varia????o do spread. Al??m do rating, os resultados indicam que os principais fatores que determinam o spread das emiss??es s??o: garantias, experi??ncia do emissor, maturidade, volume, prest??gio do banco coordenador, benef??cios fiscais e cen??rio econ??mico. Em raz??o dos resultados alcan??ados, foram abordados temas relacionados ?? efici??ncia de mercado, tais como conflitos de ag??ncia, assimetria de informa????o e sele????o adversa. Os resultados demonstram que, as emiss??es com garantia remuneraram seus investidores com spread maior que as sem garantia, e esse pr??mio variou de 35 a 38 basis points. Os resultados foram interpretados no contexto da teoria da ag??ncia e se assemelham aos encontrados por John, Lynch e Puri (2003) para o mercado norte-americano de corporatebonds. Finalmente, o cen??rio econ??mico favor??vel, medido pelo ??ndice Emerging Market Bond Index - Brazil (EMBI+ BR), demonstrou-senegativamentecorrelacionado com o spread, e esses resultados foram interpretados como efeitos de assimetria de informa????o e sele????o adversa presentes no mercado local de emiss??o de d??vida.

Made available in DSpace on 2014-12-17T14:54:57Z (GMT). No. of bitstreams: 1 RaquelEPE_TESE_1-80.pdf: 4159949 bytes, checksum: 681cbd9cd3971c6a6645b13d18e3ce70 (MD5) Previous issue date: 2009-03-20 / Conselho Nacional de Desenvolvimento Cient?fico e Tecnol?gico / With the rapid growth of databases of various types (text, multimedia, etc..), There exist a need to propose methods for ordering, access and retrieve data in a simple and fast way. The images databases, in addition to these needs, require a representation of the images so that the semantic content characteristics are considered. Accordingly, several proposals such as the textual annotations based retrieval has been made. In the annotations approach, the recovery is based on the comparison between the textual description that a user can make of images and descriptions of the images stored in database. Among its drawbacks, it is noted that the textual description is very dependent on the observer, in addition to the computational effort required to describe all the images in database. Another approach is the content based image retrieval - CBIR, where each image is represented by low-level features such as: color, shape, texture, etc. In this sense, the results in the area of CBIR has been very promising. However, the representation of the images semantic by low-level features is an open problem. New algorithms for the extraction of features as well as new methods of indexing have been proposed in the literature. However, these algorithms become increasingly complex. So, doing an analysis, it is natural to ask whether there is a relationship between semantics and low-level features extracted in an image? and if there is a relationship, which descriptors better represent the semantic? which leads us to a new question: how to use descriptors to represent the content of the images?. The work presented in this thesis, proposes a method to analyze the relationship between low-level descriptors and semantics in an attempt to answer the questions before. Still, it was observed that there are three possibilities of indexing images: Using composed characteristic vectors, using parallel and independent index structures (for each descriptor or set of them) and using characteristic vectors sorted in sequential order. Thus, the first two forms have been widely studied and applied in literature, but there were no records of the third way has even been explored. So this thesis also proposes to index using a sequential structure of descriptors and also the order of these descriptors should be based on the relationship that exists between each descriptor and semantics of the users. Finally, the proposed index in this thesis revealed better than the traditional approachs and yet, was showed experimentally that the order in this sequence is important and there is a direct relationship between this order and the relationship of low-level descriptors with the semantics of the users / Na recupera??o de imagens basada no conte?do - CBIR, cada imagem ? representada pelas suas caracter?sticas de baixo n?vel como s?o: cor, forma, textura, etc. A representa??o da sem?ntica das imagens por caracter?sticas de baixo n?vel ? um problema em aberto. Novos algoritmos para a extra??o de caracter?sticas assim como novos m?todos de indexa??o tem sido propostos na literatura. Por?m, estes algoritmos tornam-se cada vez mais complexos surgindo assim uma serie de questionamentos, tais como: existe uma rela??o entre a sem?ntica e as caracter?sticas de baixo n?vel extra?das em uma imagem? quais descritores representam melhor esta sem?ntica? responder estes questionamentos nos leva a um novo: quantos descritores usar para a representa??o do conte?do das imagens?. Nesta tese propomos um m?todo para analisar a rela??o que existe entre descritores de baixo n?vel e a sem?ntica, na tentativa de responder os questionamentos formulados. Ainda, propoe-se uma indexa??o dos vetores de caracter?sticas ordenados de forma seq?encial, a qual foi comparada com as formas de indexa??o tradicionais. Assim, para indexar as imagens usando uma estrutura seq?encial dos descritores, foi estabelecido uma ordem segundo a rela??o que existe entre cada descritor e a sem?ntica das imagens. Finalmente, a proposta de indexa??o realizada nesta tese mostrou-se superior ?s propostas tradicionais e ainda, mostrou-se experimentalmente que a ordem nesta seq??ncia ? relevante e existe uma rela??o direta entre esta ordem e a rela??o dos descritores de baixo n?vel com a sem?ntica das imagens. Como estrutura de indexa??o foi usada uma rede TS-SL-SOM e ? proposta um novo algoritmo de treinamento nesta rede de forma que a efici?ncia alcan?ada seja otimizada. Finalmente, para poder estabelecer o grau de sem?ntica extra?da por cada descritor s?o propostos algoritmos e ?ndices que quantificam esta sem?ntica de tal forma que os descritores sejam compar?veis e se consiga escolher quais descritores usar segundo o problema dado