Perceptron and SVM learning with generalized cost models

Research output: Journal contributionsJournal articlesResearchpeer-review

Standard

Perceptron and SVM learning with generalized cost models. / Geibel, Peter; Brefeld, Ulf; Wysotzki, Fritz.
In: Intelligent Data Analysis, Vol. 8, No. 5, 2004, p. 439-455.

Research output: Journal contributionsJournal articlesResearchpeer-review

Harvard

APA

Vancouver

Geibel P, Brefeld U, Wysotzki F. Perceptron and SVM learning with generalized cost models. Intelligent Data Analysis. 2004;8(5):439-455. doi: 10.1137/S1052623499362111

Bibtex

@article{5cdd14205152422aa7f5aa2e06e8f97e,
title = "Perceptron and SVM learning with generalized cost models",
abstract = "Learning algorithms from the fields of artificial neural networks and machine learning, typically, do not take any costs into account or allow only costs depending on the classes of the examples that are used for learning. As an extension of class dependent costs, we consider costs that are example, i.e. feature and class dependent. We derive a cost-sensitive perceptron learning rule for non-separable classes, that can be extended to multi-modal classes (DIPOL) and present a natural cost-sensitive extension of the support vector machine (SVM). We also derive an approach for including example dependent costs into an arbitrary cost-insensitive learning algorithm by sampling according to modified probability distributions.",
keywords = "Informatics, Business informatics",
author = "Peter Geibel and Ulf Brefeld and Fritz Wysotzki",
year = "2004",
doi = "10.1137/S1052623499362111",
language = "English",
volume = "8",
pages = "439--455",
journal = "Intelligent Data Analysis",
issn = "1088-467X",
publisher = "IOS Press BV",
number = "5",

}

RIS

TY - JOUR

T1 - Perceptron and SVM learning with generalized cost models

AU - Geibel, Peter

AU - Brefeld, Ulf

AU - Wysotzki, Fritz

PY - 2004

Y1 - 2004

N2 - Learning algorithms from the fields of artificial neural networks and machine learning, typically, do not take any costs into account or allow only costs depending on the classes of the examples that are used for learning. As an extension of class dependent costs, we consider costs that are example, i.e. feature and class dependent. We derive a cost-sensitive perceptron learning rule for non-separable classes, that can be extended to multi-modal classes (DIPOL) and present a natural cost-sensitive extension of the support vector machine (SVM). We also derive an approach for including example dependent costs into an arbitrary cost-insensitive learning algorithm by sampling according to modified probability distributions.

AB - Learning algorithms from the fields of artificial neural networks and machine learning, typically, do not take any costs into account or allow only costs depending on the classes of the examples that are used for learning. As an extension of class dependent costs, we consider costs that are example, i.e. feature and class dependent. We derive a cost-sensitive perceptron learning rule for non-separable classes, that can be extended to multi-modal classes (DIPOL) and present a natural cost-sensitive extension of the support vector machine (SVM). We also derive an approach for including example dependent costs into an arbitrary cost-insensitive learning algorithm by sampling according to modified probability distributions.

KW - Informatics

KW - Business informatics

UR - http://www.scopus.com/inward/record.url?scp=56449118713&partnerID=8YFLogxK

U2 - 10.1137/S1052623499362111

DO - 10.1137/S1052623499362111

M3 - Journal articles

AN - SCOPUS:56449118713

VL - 8

SP - 439

EP - 455

JO - Intelligent Data Analysis

JF - Intelligent Data Analysis

SN - 1088-467X

IS - 5

ER -

Documents

DOI