TY - GEN
T1 - Learning through changes
T2 - 5th IEEE International Conference on Data Mining, ICDM 2005
AU - Zhang, Kun
AU - Xu, Zujia
AU - Peng, Jing
AU - Buckles, Bill
PY - 2005
Y1 - 2005
N2 - In practice, learning from data is often hampered by the limited training examples. In this paper, as the size of training data varies, we empirically investigate several probability estimation tree algorithms over eighteen binary classification problems. Nine metrics are used to evaluate their performances. Our aggregated results show that ensemble trees consistently outperform single trees. Confusion factor trees(CFT) register poor calibration even as training size increases, which shows that CFTs are potentially biased if data sets have small noise. We also provide analysis on the observed performance of the tree algorithms.
AB - In practice, learning from data is often hampered by the limited training examples. In this paper, as the size of training data varies, we empirically investigate several probability estimation tree algorithms over eighteen binary classification problems. Nine metrics are used to evaluate their performances. Our aggregated results show that ensemble trees consistently outperform single trees. Confusion factor trees(CFT) register poor calibration even as training size increases, which shows that CFTs are potentially biased if data sets have small noise. We also provide analysis on the observed performance of the tree algorithms.
UR - http://www.scopus.com/inward/record.url?scp=33749570764&partnerID=8YFLogxK
U2 - 10.1109/ICDM.2005.88
DO - 10.1109/ICDM.2005.88
M3 - Conference contribution
AN - SCOPUS:33749570764
SN - 0769522785
SN - 9780769522784
T3 - Proceedings - IEEE International Conference on Data Mining, ICDM
SP - 817
EP - 820
BT - Proceedings - Fifth IEEE International Conference on Data Mining, ICDM 2005
Y2 - 27 November 2005 through 30 November 2005
ER -