Title :
Pool-based active learning based on incremental decision tree
Author :
Wang, Shuo ; Wang, Jian-jian ; Gao, Xiang-hui ; Wang, Xue-zheng
Author_Institution :
Machine Learning Center, Hebei Univ., Baoding, China
Abstract :
The pool-based active learning intends to collect the samples into the pool firstly, and selects the best informative sample from it which has no label to add into the training sets for updating the classifier secondly. This paper proposed a new method based on the incremental decision tree algorithm to measure the ambiguity of the unlabeled samples for the sample selection in the active learning.
Keywords :
decision trees; learning (artificial intelligence); pattern classification; classifier; incremental decision tree algorithm; pool-based active learning; sample selection; training sets; Accuracy; Algorithm design and analysis; Classification algorithms; Decision trees; Machine learning; Testing; Training; Ambiguity; Incremental decision tree; Pool-based active learning; Sample selection; Unlabeled samples;
Conference_Titel :
Machine Learning and Cybernetics (ICMLC), 2010 International Conference on
Conference_Location :
Qingdao
Print_ISBN :
978-1-4244-6526-2
DOI :
10.1109/ICMLC.2010.5581052
