arrow
返回

Wrapper feature selection for small sample size data driven by complete error estimates

delete2012-10-01
delete14
PRE
AI
M
Martin Macaš *
L
Lenka Lhotská
E
Eduard Bakštein
D
Daniel Novák
J
Jiří Wild
T
Tomáš Sieger
P
Pavel Vostatek
R
Robert Jech
DOI:10.1016/j.cmpb.2012.02.006delete
delete原文链接
delete分享
delete收藏
摘要

摘要

En 中文
This paper focuses on wrapper-based feature selection for a 1-nearest neighbor classifier. we consider in particular the case of a small sample size with a few hundred instances, which is common in biomedical applications. We propose a technique for calculating the complete bootstrap for a 1-nearest-neighbor classifier (i.e., averaging over all desired test/train partitions of the data). The complete bootstrap and the complete cross-validation error estimate with lower variance are applied as novel selection criteria and are compared with the standard bootstrap and cross-validation in combination with three optimization techniques sequential forward selection (SFS), binary particle swarm optimization (BPSO) and simplified social impact theory based optimization (SSITO). The experimental comparison based on ten datasets draws the following conclusions: for all three search methods examined here, the complete criteria are a significantly better choice than standard 2-fold cross-validation, 10-fold cross-validation and bootstrap with 50 trials irrespective of the selected output number of iterations. All the complete criterion-based 1NN wrappers with SFS search performed better than the widely-used FILTER and SIMBA methods. We also demonstrate the benefits and properties of our approaches on an important and novel real-world application of automatic detection of the subthalamic nucleus. (c) 2012 Elsevier Ireland Ltd. All rights reserved.
Keyword:
Small sample size
Feature selection
Wrappers
Cross-validation
Bootstrap
Nearest neighbor
Subthalamic nucleus detection
Swarm intelligence
AI总结

AI总结

对已上传原文的论文进行重点信息的提取,主要内容包括:简要概述、研究摘要、背景介绍、关键亮点、图文解析、展望与总结。

期刊

Computer Methods and Programs in Biomedicine 封面图
Computer Methods and Programs in Biomedicine
IF:
4.8
论文数:
7.0K
被引数:
2.6W

机构

C
czech technical university prague
学者数:
6.6K
论文数: 5.3K
被引数: 3.8K
C
Charles University Prague
学者数:
3.0W
论文数: 2.2W
被引数: 2.4W
引用论文

引用论文

Electrocardiographic Criteria for ST-Elevation Myocardial Infarction in Patients With Left Ventricular Hypertrophy
err2012-10-01
err0
PREAI
errEhrin J. Armstrong; Ameya R. Kulkarni; Prashant D. Bhave; Kurt S. Hoffmayer; John S. MacGregor; John C. Stein; Scott Kinlay; Peter Ganz; James M. McCabe
err分享
err收藏
Kinds of Reasons
err
IF0
err2010-03-25
err0
PREAI
errMaria Alvarez
err分享
err收藏
Analysis of Phagosomal Antigen Degradation by Flow Organellocytometry
err2016-01-01
err0
errOAAI
errEik Hoffmann; Anne-Marie Pauwels; Andrés Alloatti; Fiorella Kotsias; Sebastian Amigorena
err分享
err收藏
Cross-validation and bootstrapping are unreliable in small sample classification
err2008-10-01
err151
errOAAI
errIsaksson, A.; Wallman, M.; Goransson, H.; Gustafsson, M. G.
err分享
err收藏
err分享
err收藏
学者 查看更多内容