RI UFLA (Universidade Federal de Lavras) >
Revistas UFLA >

Please use this identifier to cite or link to this item: http://repositorio.ufla.br/jspui/handle/1/14530

???metadata.dc.creator???: Saito, Nathália Suemi
Arguello, Fernanda Viana Paiva
Moreira, Maurício Alves
Santos, Alexandre Rosa dos
Eugenio, Fernando Coelho
Figueiredo, Alvaro Costa
Keywords: Data Mining; Landscape ecology; GeoDMA
Publisher: CERNE
???metadata.dc.date???: 29-Apr-2016
Other Identifiers: http://www.cerne.ufla.br/site/index.php/CERNE/article/view/1150
Description: The landscape ecology metrics associated with data mining can be used to increase the potential of remote sensing data analysis and applications, being an important tool for decision making. The present study aimed to use data mining techniques and landscape ecology metrics to classify and quantify  different types of vegetation using a multitemporal analysis (2001 and 2011), in São Luís do Paraitinga city, São Paulo, Brazil. Object-based image analyses and the C4.5 data-mining algorithm were used for automated classification. Classification accuracies were assessed using the kappa index of agreement and the recently proposed allocation and quantity disagreement measures. Four land use and land cover classes were mapped, including Eucalyptus plantations, whose area increased from 4.4% to 8.6%. The automatic classification showed a kappa index of 0.79 and 0.80, quantity disagreements of 2% e 3.5% and allocation measures of 5.5% and 5% for 2001 and 2011, respectively. We therefore concluded that the data mining method and landscape ecology metrics were efficient in separating vegetation classes.
???metadata.dc.language???: eng
Appears in Collections:CERNE

Files in This Item:

There are no files associated with this item.

Items in DSpace are protected by copyright, with all rights reserved, unless otherwise indicated.

View Statistics


DSpace Software Copyright © 2002-2010  Duraspace - Feedback