Multiscale Bagging and Its Applications

Hidetoshi SHIMODAIRA; Takafumi KANAMORI; Masayoshi AOKI; Kouta MINE

doi:10.1587/transinf.E94.D.1924

Multiscale Bagging and Its Applications

Hidetoshi SHIMODAIRA, Takafumi KANAMORI, Masayoshi AOKI, Kouta MINE

Full Text Views

0

Cite this

Summary :

We propose multiscale bagging as a modification of the bagging procedure. In ordinary bagging, the bootstrap resampling is used for generating bootstrap samples. We replace it with the multiscale bootstrap algorithm. In multiscale bagging, the sample size m of bootstrap samples may be altered from the sample size n of learning dataset. For assessing the output of a classifier, we compute bootstrap probability of class label; the frequency of observing a specified class label in the outputs of classifiers learned from bootstrap samples. A scaling-law of bootstrap probability with respect to σ²=n/m has been developed in connection with the geometrical theory. We consider two different ways for using multiscale bagging of classifiers. The first usage is to construct a confidence set of class labels, instead of a single label. The second usage is to find inputs close to decision boundaries in the context of query by bagging for active learning. It turned out, interestingly, that an appropriate choice of m is m =-n, i.e., σ²=-1, for the first usage, and m =∞, i.e., σ²=0, for the second usage.

Publication: IEICE TRANSACTIONS on Information Vol.E94-D No.10 pp.1924-1932

Publication Date: 2011/10/01

Publicized

Online ISSN: 1745-1361

DOI: 10.1587/transinf.E94.D.1924

Type of Manuscript: Special Section PAPER (Special Section on Information-Based Induction Sciences and Machine Learning)

Category

Cite this

Copy

Hidetoshi SHIMODAIRA, Takafumi KANAMORI, Masayoshi AOKI, Kouta MINE, "Multiscale Bagging and Its Applications" in IEICE TRANSACTIONS on Information, vol. E94-D, no. 10, pp. 1924-1932, October 2011, doi: 10.1587/transinf.E94.D.1924.
Abstract: We propose multiscale bagging as a modification of the bagging procedure. In ordinary bagging, the bootstrap resampling is used for generating bootstrap samples. We replace it with the multiscale bootstrap algorithm. In multiscale bagging, the sample size m of bootstrap samples may be altered from the sample size n of learning dataset. For assessing the output of a classifier, we compute bootstrap probability of class label; the frequency of observing a specified class label in the outputs of classifiers learned from bootstrap samples. A scaling-law of bootstrap probability with respect to σ²=n/m has been developed in connection with the geometrical theory. We consider two different ways for using multiscale bagging of classifiers. The first usage is to construct a confidence set of class labels, instead of a single label. The second usage is to find inputs close to decision boundaries in the context of query by bagging for active learning. It turned out, interestingly, that an appropriate choice of m is m =-n, i.e., σ²=-1, for the first usage, and m =∞, i.e., σ²=0, for the second usage.
URL: https://global.ieice.org/en_transactions/information/10.1587/transinf.E94.D.1924/_p

Copy

@ARTICLE{e94-d_10_1924,
author={Hidetoshi SHIMODAIRA, Takafumi KANAMORI, Masayoshi AOKI, Kouta MINE, },
journal={IEICE TRANSACTIONS on Information},
title={Multiscale Bagging and Its Applications},
year={2011},
volume={E94-D},
number={10},
pages={1924-1932},
abstract={We propose multiscale bagging as a modification of the bagging procedure. In ordinary bagging, the bootstrap resampling is used for generating bootstrap samples. We replace it with the multiscale bootstrap algorithm. In multiscale bagging, the sample size m of bootstrap samples may be altered from the sample size n of learning dataset. For assessing the output of a classifier, we compute bootstrap probability of class label; the frequency of observing a specified class label in the outputs of classifiers learned from bootstrap samples. A scaling-law of bootstrap probability with respect to σ²=n/m has been developed in connection with the geometrical theory. We consider two different ways for using multiscale bagging of classifiers. The first usage is to construct a confidence set of class labels, instead of a single label. The second usage is to find inputs close to decision boundaries in the context of query by bagging for active learning. It turned out, interestingly, that an appropriate choice of m is m =-n, i.e., σ²=-1, for the first usage, and m =∞, i.e., σ²=0, for the second usage.},
keywords={},
doi={10.1587/transinf.E94.D.1924},
ISSN={1745-1361},
month={October},}

Copy

TY - JOUR
TI - Multiscale Bagging and Its Applications
T2 - IEICE TRANSACTIONS on Information
SP - 1924
EP - 1932
AU - Hidetoshi SHIMODAIRA
AU - Takafumi KANAMORI
AU - Masayoshi AOKI
AU - Kouta MINE
PY - 2011
DO - 10.1587/transinf.E94.D.1924
JO - IEICE TRANSACTIONS on Information
SN - 1745-1361
VL - E94-D
IS - 10
JA - IEICE TRANSACTIONS on Information
Y1 - October 2011
AB - We propose multiscale bagging as a modification of the bagging procedure. In ordinary bagging, the bootstrap resampling is used for generating bootstrap samples. We replace it with the multiscale bootstrap algorithm. In multiscale bagging, the sample size m of bootstrap samples may be altered from the sample size n of learning dataset. For assessing the output of a classifier, we compute bootstrap probability of class label; the frequency of observing a specified class label in the outputs of classifiers learned from bootstrap samples. A scaling-law of bootstrap probability with respect to σ²=n/m has been developed in connection with the geometrical theory. We consider two different ways for using multiscale bagging of classifiers. The first usage is to construct a confidence set of class labels, instead of a single label. The second usage is to find inputs close to decision boundaries in the context of query by bagging for active learning. It turned out, interestingly, that an appropriate choice of m is m =-n, i.e., σ²=-1, for the first usage, and m =∞, i.e., σ²=0, for the second usage.
ER -