DSpace Coleção:http://repositorio.ufgd.edu.br/jspui/handle/prefix/17382021-12-04T09:08:23Z2021-12-04T09:08:23ZFerramenta para extrair de coordenadas reais de imagens digitaisLopes, Rafael Ernestohttp://repositorio.ufgd.edu.br/jspui/handle/prefix/46362021-12-01T01:42:16Z2021-05-12T00:00:00ZTítulo: Ferramenta para extrair de coordenadas reais de imagens digitais
Autor(es): Lopes, Rafael Ernesto
Primeiro Orientador: Santos, Wellington Lima dos
Abstract: In this work, the objective was to develop a digitizer, in the form of web application, to extract world coordinates from images of XY-type graphs with Cartesian axes represented in linear scales. The analysis of 13 digitizing tools served as a basis for the implementation of some functionalities such as the display of statistical parameters associated with the definition of scale, as well as improvements in the definition of control points, which contributes to the configuration of more reliable scales and, consequently, more results accurate. The mathematical model used to adjust the scale equations was the same to perform linear geometric transformations affine in the plane and, to validate it, a unit radius circumference contained in an image transformed by shear and then by rotation was digitized. The statistical analysis showed excellent results, which allow us to conclude that the mathematical model was used correctly.
Editor: Universidade Federal da Grande Dourados
Tipo: Trabalho de Conclusão de Curso2021-05-12T00:00:00ZBolsa de valores: plataformas computacionais e robôsMartins, Carlos Antônio Gonçalveshttp://repositorio.ufgd.edu.br/jspui/handle/prefix/46352021-12-01T01:41:08Z2021-05-10T00:00:00ZTítulo: Bolsa de valores: plataformas computacionais e robôs
Autor(es): Martins, Carlos Antônio Gonçalves
Primeiro Orientador: Oeiras Lachi, Janne Yukiko Yoshikawa
Abstract: The present work explains how trading platforms and robots operate on the stock exchange as instruments for executing investor strategies, presenting the various concepts present in each of the stages for the process of trading financial assets and pointing out how and which changes are created from the use of these technologies. For this purpose, bibliographic research will be used to carry out a general study on the topic through books, articles and news portals on technology and economics.
Editor: Universidade Federal da Grande Dourados
Tipo: Trabalho de Conclusão de Curso2021-05-10T00:00:00ZCode for Pantanal: desenvolvimento de tecnologias cívicas para o Mato Grosso do SulDionisio, Antonio NetoSouza, Paulo Henrique Nascimento dehttp://repositorio.ufgd.edu.br/jspui/handle/prefix/46342021-12-01T01:39:39Z2021-05-12T00:00:00ZTítulo: Code for Pantanal: desenvolvimento de tecnologias cívicas para o Mato Grosso do Sul
Autor(es): Dionisio, Antonio Neto; Souza, Paulo Henrique Nascimento de
Primeiro Orientador: Carbone, Felipe José
Abstract: Nowadays, digital technologies and internet access are part of most people's lives. The expansion of the internet combined with new development tools in programming has allowed the community development of applications to solve the most varied urban problems. This work describes the implementation of a web platform for the dissemination of applications based on civic technologies, through open-source community development in Mato Grosso do Sul, entitled “Code for Pantanal” project. Such a platform is being conceived through the use of open source tools, and the back-end is with a NodeJS API using the ExpressJS framework that is linking the database in MongoDB, reporting Json to the front-end in ReactJS. Among the functionalities developed are: i) visualization of registered projects, ii) request to participate in a project, iii) indication for new projects, iv) summary of project documentation, v) information about the team, vi) didactic material, including tutorials and video lessons on projects, sharing source code and first steps in programming. The project, in addition to educating the authors' training in technology, proves to be of great contribution to society, since there is the possibility of this project shortening the relationship between the community and municipal entities, accelerating the response to solve urban problems.
Editor: Universidade Federal da Grande Dourados
Tipo: Trabalho de Conclusão de Curso2021-05-12T00:00:00ZRedução de template aplicado a K-Nearest Neighbor e Floresta de Caminhos Ótimos em problemas de reconhecimento de padrõesLemke, Francisco SamuelThiry Neto, Paulo Albertohttp://repositorio.ufgd.edu.br/jspui/handle/prefix/43382020-10-22T13:55:05Z2013-04-09T00:00:00ZTítulo: Redução de template aplicado a K-Nearest Neighbor e Floresta de Caminhos Ótimos em problemas de reconhecimento de padrões
Autor(es): Lemke, Francisco Samuel; Thiry Neto, Paulo Alberto
Primeiro Orientador: Amorim, Willian Paraguassu
Abstract: An of the classic algorithms in the area of learning and classication of data is the KNN (K Nearest Neighbors). The KNN algorithm is supervised learning, where the consultation result of a new instance is classied based on the dominant class of K nearest neighbors. The KNN has some disadvantages, as each new consultation will need to calculate the distance to all training samples to reach a certain ranking, requiring a high processing time. Another approach that has been used in pattern classication is the OPF (Optimum Path-Forest). OPF is a framework for data classication, supervised and unsupervised, particularly eective in image classication, which reduces problems of pattern classication to a problem of partitioning the vertices of a graph. One of the possible problems that may be encountered in the use of the technique OPF is the classication of large volumes of data. For the training phase is necessary to calculate the complete graph, in other words, the distance between all samples, which in a training base too large will cause a high processing time. The method Template Reduction (TR), the present paper aims to obtain a smaller model, a subset extracted from the training set. The main idea is to nd the TR border regions between dierent classes of the training set, and discard prototypes that are far from this region. In this study, we evaluated the use of TR, and after applying the KNN and OPF. We evaluate the results, the proposal TR, achieved satisfactory results, while not getting the best results of classication rate, yet is ecient, signicantly reducing the volume of data, and accelerating the learning process and classication.
Editor: Universidade Federal da Grande Dourados
Tipo: Trabalho de Conclusão de Curso2013-04-09T00:00:00Z