Bag Dissimilarities for Multiple Instance Learning

被引：0

作者：

Tax, David M. J. ^{[1
]}

Loog, Marco ^{[1
]}

Duin, Robert P. W. ^{[1
]}

Cheplygina, Veronika ^{[1
]}

Lee, Wan-Jui ^{[1
]}

机构：

[1] Delft Univ Technol, Pattern Recognit Lab, NL-2628 CD Delft, Netherlands

来源：

SIMILARITY-BASED PATTERN RECOGNITION: FIRST INTERNATIONAL WORKSHOP, SIMBAD 2011 | 2011年 / 7005卷

关键词：

pattern recognition; multiple instance learning; dissimilarity representation;

D O I：

暂无

中图分类号：

TP18 [人工智能理论];

学科分类号：

081104 ; 0812 ; 0835 ; 1405 ;

摘要：

When objects cannot be represented well by single feature vectors, a collection of feature vectors can be used. This is what is done in Multiple Instance learning, where it is called a bag of instances. By using a bag of instances, an object gains more internal structure than when a single feature vector is used. This improves the expressiveness of the representation, but also adds complexity to the classification of the object. This paper shows that for the situation that not a single instance determines the class label of a bag, simple bag dissimilarity measures can significantly outperform standard multiple instance classifiers. In particular a measure that computes just the average minimum distance between instances, or a measure that uses the Earth Mover's distance, perform very well.

引用

页码：222 / 234

页数：13