dc.contributor | Fidelis, Marcos Vinicius | |
dc.contributor | Fidelis, Marcos Vinicius | |
dc.contributor | Borges, Helyane Bronoski | |
dc.contributor | Almeida, Simone de | |
dc.creator | Ferreira, André Luis | |
dc.date.accessioned | 2020-11-19T21:04:55Z | |
dc.date.accessioned | 2022-12-06T15:14:20Z | |
dc.date.available | 2020-11-19T21:04:55Z | |
dc.date.available | 2022-12-06T15:14:20Z | |
dc.date.created | 2020-11-19T21:04:55Z | |
dc.date.issued | 2012-06-05 | |
dc.identifier | FERREIRA, André Luís. Mapas auto organizáveis na descoberta e validação de padrões em bases de dados. 2012. 74 f. Trabalho de Conclusão do Curso (Graduação) - Universidade Tecnológica Federal do Paraná, Ponta Grossa, 2012. | |
dc.identifier | http://repositorio.utfpr.edu.br/jspui/handle/1/16755 | |
dc.identifier.uri | https://repositorioslatinoamericanos.uchile.cl/handle/2250/5262830 | |
dc.description.abstract | The Self-Organizing Maps (SOM) of Kohonen is an excellent and powerful algorithm for data mining based in Artificial Neural Networks (ANN). Through the SOM is possible to do processing and analysis of databases of large size, and then discover/retrieve the items in the data groups present in the database, having the items in each group that share similar properties in common. From the clusters formed by SOM, is possible then perform an focused analysis of the resulting clusters. In some cases however, the SOM find groups of data not classified or understandable to the context, since it groups the data based on all of its n characteristics. For the results of clustering being/get valid and useful information, is important that the data be pre-processed to suit the algorithm, and subsequently the classification performed by the SOM be/is/get validated in order to obtain the precision with which the SOM classified the data. With the standardization of the data and subsequent validation of the clusters is then possible to verify the reliability of the algorithm. From the data resulting from all these processes it is now possible to extract this data in a humanly understandable way, providing information that previously would not be understandable or predictable in the raw data, favoring the interpretation of results. | |
dc.publisher | Universidade Tecnológica Federal do Paraná | |
dc.publisher | Ponta Grossa | |
dc.publisher | Brasil | |
dc.publisher | Departamento de Informática | |
dc.publisher | Tecnologia em Análise e Desenvolvimento de Sistemas | |
dc.publisher | UTFPR | |
dc.rights | openAccess | |
dc.subject | Banco de dados | |
dc.subject | Mineração de dados (Computação) | |
dc.subject | Redes neurais (Computação) | |
dc.subject | Data bases | |
dc.subject | Data mining | |
dc.subject | Neural networks (Computer science) | |
dc.title | Mapas auto organizáveis na descoberta e validação de padrões em bases de dados | |
dc.type | bachelorThesis | |