A New SVM Multiclass Incremental Learning Algorithm
A new support vector machine (SVM) multiclass incremental learning algorithm is proposed. To each class training sample, the hyperellipsoidal classifier that includes as many samples as possible and pushes the outlier samples away is trained in the feature space. When the new samples are added to th...
Main Authors: | , , |
---|---|
Format: | Article |
Language: | English |
Published: |
Hindawi Limited
2015-01-01
|
Series: | Mathematical Problems in Engineering |
Online Access: | http://dx.doi.org/10.1155/2015/745815 |
id |
doaj-b5fceec3d7f94683918139c69d90b106 |
---|---|
record_format |
Article |
spelling |
doaj-b5fceec3d7f94683918139c69d90b1062020-11-24T22:23:55ZengHindawi LimitedMathematical Problems in Engineering1024-123X1563-51472015-01-01201510.1155/2015/745815745815A New SVM Multiclass Incremental Learning AlgorithmYuping Qin0Dan Li1Aihua Zhang2College of Engineering, Bohai University, Jinzhou 121013, ChinaCollege of Mathematics and Physics, Bohai University, Jinzhou 121013, ChinaCollege of Engineering, Bohai University, Jinzhou 121013, ChinaA new support vector machine (SVM) multiclass incremental learning algorithm is proposed. To each class training sample, the hyperellipsoidal classifier that includes as many samples as possible and pushes the outlier samples away is trained in the feature space. When the new samples are added to the classification system, the algorithm reuses the old classifiers that have nothing to do with the new sample classes. To be classified sample, the Mahalanobis distances are used to decide the class of classified sample. If the sample point is not surrounded by any hyperellipsoidal or is surrounded by more than one hyperellipsoidal, the membership is used to confirm its class. The experimental results show that the algorithm has higher performance in classification precision and classification speed.http://dx.doi.org/10.1155/2015/745815 |
collection |
DOAJ |
language |
English |
format |
Article |
sources |
DOAJ |
author |
Yuping Qin Dan Li Aihua Zhang |
spellingShingle |
Yuping Qin Dan Li Aihua Zhang A New SVM Multiclass Incremental Learning Algorithm Mathematical Problems in Engineering |
author_facet |
Yuping Qin Dan Li Aihua Zhang |
author_sort |
Yuping Qin |
title |
A New SVM Multiclass Incremental Learning Algorithm |
title_short |
A New SVM Multiclass Incremental Learning Algorithm |
title_full |
A New SVM Multiclass Incremental Learning Algorithm |
title_fullStr |
A New SVM Multiclass Incremental Learning Algorithm |
title_full_unstemmed |
A New SVM Multiclass Incremental Learning Algorithm |
title_sort |
new svm multiclass incremental learning algorithm |
publisher |
Hindawi Limited |
series |
Mathematical Problems in Engineering |
issn |
1024-123X 1563-5147 |
publishDate |
2015-01-01 |
description |
A new support vector machine (SVM) multiclass incremental learning algorithm is proposed. To each class training sample, the hyperellipsoidal classifier that includes as many samples as possible and pushes the outlier samples away is trained in the feature space. When the new samples are added to the classification system, the algorithm reuses the old classifiers that have nothing to do with the new sample classes. To be classified sample, the Mahalanobis distances are used to decide the class of classified sample. If the sample point is not surrounded by any hyperellipsoidal or is surrounded by more than one hyperellipsoidal, the membership is used to confirm its class. The experimental results show that the algorithm has higher performance in classification precision and classification speed. |
url |
http://dx.doi.org/10.1155/2015/745815 |
work_keys_str_mv |
AT yupingqin anewsvmmulticlassincrementallearningalgorithm AT danli anewsvmmulticlassincrementallearningalgorithm AT aihuazhang anewsvmmulticlassincrementallearningalgorithm AT yupingqin newsvmmulticlassincrementallearningalgorithm AT danli newsvmmulticlassincrementallearningalgorithm AT aihuazhang newsvmmulticlassincrementallearningalgorithm |
_version_ |
1725763326698323968 |