Int. Conf. Artificial Intelligence ICAI 2009 , 2009-08-14

Title : ( Improved DAG SVM: A New Method for Multi-Class SVM Classification )

Authors: Mostafa Sabzekar , Mohammad GhasemiGol , Mahmoud Naghibzadeh , Hadi Sadoghi Yazdi ,

Citation: BibTeX | EndNote

Abstract

In this paper, we present our method which is a performance improvement to the Directed Acyclic Graph Support Vector Machines (DAG SVM). It suggests a weighted multi-class classification technique which divides the input space into several subspaces. In the training phase of the technique, for each subspace, a DAG SVM is trained and its probability density function (pdf) is guesstimated. In the test phase, fit in value of each input pattern to every subspace is calculated using the pdf of the subspace as the weight of each DAG SVM. Finally, a fusion operation is defined and applied to the DAG SVM outputs to decide the class label of the given input pattern. Evaluation results show the prominence of our method of multi-class classification compared with DAG SVM. Some data sets including synthetic one, the iris, and the wine data sets relative standard DAG SVM, were used for the evaluation.

Keywords

, DAG SVM, classifier combination, multi-class classification.
برای دانلود از شناسه و رمز عبور پرتال پویا استفاده کنید.

@inproceedings{paperid:1015403,
author = {Sabzekar, Mostafa and GhasemiGol, Mohammad and Naghibzadeh, Mahmoud and Sadoghi Yazdi, Hadi},
title = {Improved DAG SVM: A New Method for Multi-Class SVM Classification},
booktitle = {Int. Conf. Artificial Intelligence ICAI 2009},
year = {2009},
keywords = {DAG SVM; classifier combination; multi-class classification.},
}

[Download]

%0 Conference Proceedings
%T Improved DAG SVM: A New Method for Multi-Class SVM Classification
%A Sabzekar, Mostafa
%A GhasemiGol, Mohammad
%A Naghibzadeh, Mahmoud
%A Sadoghi Yazdi, Hadi
%J Int. Conf. Artificial Intelligence ICAI 2009
%D 2009

[Download]