DocumentCode :
2423872
Title :
A Study on Class Noise Detection and Elimination
Author :
Garcia, Luís Paulo F ; Lorena, Ana Carolina ; Carvalho, André C P L F
Author_Institution :
Univ. de Sao Paulo (USP), Sao Carlos, Brazil
fYear :
2012
fDate :
20-25 Oct. 2012
Firstpage :
13
Lastpage :
18
Abstract :
Real data may present a significant amount of noise, generated by inaccuracies in data collection, transmission and storage. The presence of noisy data in a training dataset used for the induction of a Machine Learning model may increase the training time and the complexity of the induced model, resulting in the deterioration of its predictive performance for new data. Noise may be found in the input and target attributes. In this study, we are concerned with noise in the class label of the target attribute. For such, we propose and experimentally investigate some simple class noise detection and elimination strategies for classification problems, introducing controlled noise levels in five UCI datasets originally free of inconsistencies. The results obtained in the experiments performed show the potential of the proposed approaches.
Keywords :
learning (artificial intelligence); noise; pattern classification; UCI datasets; class noise detection; class noise elimination; classification problems; elimination strategies; machine learning model; noisy data; target attribute; training time; Accuracy; Data models; Noise; Noise level; Noise measurement; Predictive models; Training; Class Noise; Ensemble of Classifiers; Machine Learning;
fLanguage :
English
Publisher :
ieee
Conference_Titel :
Neural Networks (SBRN), 2012 Brazilian Symposium on
Conference_Location :
Curitiba
ISSN :
1522-4899
Print_ISBN :
978-1-4673-2641-4
Type :
conf
DOI :
10.1109/SBRN.2012.49
Filename :
6374817
Link To Document :
بازگشت