Title :
Bagging algorithm for pixel classification
Author :
Novakovic, Jasmina
Author_Institution :
Grad. Sch. of Comput. Sci., Megatrend Univ., Belgrade, Serbia
Abstract :
In this research, Bagging algorithm that incorporates different classifier into classifier ensembles models for pixel classification is suggested. We chose classifier ensembles with decision trees, as the base classifiers. In the problem of pixel classification, experimental results demonstrate the effectiveness of the Bagging with forest of random trees (RandomForest) as base classifier compared to other methods.
Keywords :
decision trees; image classification; Bagging algorithm; decision trees; pixel classification; random trees; Accuracy; Bagging; Classification algorithms; Decision trees; Image segmentation; Machine learning; Vegetation; Bagging; decision tree; image; pixel classification;
Conference_Titel :
Telecommunications Forum (TELFOR), 2011 19th
Conference_Location :
Belgrade
Print_ISBN :
978-1-4577-1499-3
DOI :
10.1109/TELFOR.2011.6143803