SEG-SSC: a framework based on synthetic examples generation for self-labeled semi-supervised classification
Self-labeled techniques are semi-supervised classification methods that address the shortage of labeled examples via a self-learning process based on supervised models. They progressively classify unlabeled data and use them to modify the hypothesis learned from labeled samples. Most relevant propos...
| Main Authors: | , , |
|---|---|
| Format: | Article |
| Published: |
Institute of Electrical and Electronics Engineers
2015
|
| Subjects: | |
| Online Access: | https://eprints.nottingham.ac.uk/45410/ |
| _version_ | 1848797125670862848 |
|---|---|
| author | Triguero, Isaac Garcia, Salvador Herrera, Francisco |
| author_facet | Triguero, Isaac Garcia, Salvador Herrera, Francisco |
| author_sort | Triguero, Isaac |
| building | Nottingham Research Data Repository |
| collection | Online Access |
| description | Self-labeled techniques are semi-supervised classification methods that address the shortage of labeled examples via a self-learning process based on supervised models. They progressively classify unlabeled data and use them to modify the hypothesis learned from labeled samples. Most relevant proposals are currently inspired by boosting schemes to iteratively enlarge the labeled set. Despite their effectiveness, these methods are constrained by the number of labeled examples and their distribution, which in many cases is sparse and scattered. The aim of this work is to design a framework, named SEG-SSC, to improve the classification performance of any given self-labeled method by using synthetic labeled data. These are generated via an oversampling technique and a positioning adjustment model that use both labeled and unlabeled examples as reference. Next, these examples are incorporated in the main stages of the self-labeling process. The principal aspects of the proposed framework are: (a) introducing diversity to the multiple classifiers used by using more (new) labeled data, (b) fulfilling labeled data distribution with the aid of unlabeled data, and (c) being applicable to any kind of self-labeled method. In our empirical studies, we have applied this scheme to four recent self-labeled methods, testing their capabilities with a large number of data sets. We show that this framework significantly improves the classification capabilities of self-labeled techniques. |
| first_indexed | 2025-11-14T19:58:54Z |
| format | Article |
| id | nottingham-45410 |
| institution | University of Nottingham Malaysia Campus |
| institution_category | Local University |
| last_indexed | 2025-11-14T19:58:54Z |
| publishDate | 2015 |
| publisher | Institute of Electrical and Electronics Engineers |
| recordtype | eprints |
| repository_type | Digital Repository |
| spelling | nottingham-454102020-05-04T17:05:49Z https://eprints.nottingham.ac.uk/45410/ SEG-SSC: a framework based on synthetic examples generation for self-labeled semi-supervised classification Triguero, Isaac Garcia, Salvador Herrera, Francisco Self-labeled techniques are semi-supervised classification methods that address the shortage of labeled examples via a self-learning process based on supervised models. They progressively classify unlabeled data and use them to modify the hypothesis learned from labeled samples. Most relevant proposals are currently inspired by boosting schemes to iteratively enlarge the labeled set. Despite their effectiveness, these methods are constrained by the number of labeled examples and their distribution, which in many cases is sparse and scattered. The aim of this work is to design a framework, named SEG-SSC, to improve the classification performance of any given self-labeled method by using synthetic labeled data. These are generated via an oversampling technique and a positioning adjustment model that use both labeled and unlabeled examples as reference. Next, these examples are incorporated in the main stages of the self-labeling process. The principal aspects of the proposed framework are: (a) introducing diversity to the multiple classifiers used by using more (new) labeled data, (b) fulfilling labeled data distribution with the aid of unlabeled data, and (c) being applicable to any kind of self-labeled method. In our empirical studies, we have applied this scheme to four recent self-labeled methods, testing their capabilities with a large number of data sets. We show that this framework significantly improves the classification capabilities of self-labeled techniques. Institute of Electrical and Electronics Engineers 2015-04-30 Article PeerReviewed Triguero, Isaac, Garcia, Salvador and Herrera, Francisco (2015) SEG-SSC: a framework based on synthetic examples generation for self-labeled semi-supervised classification. IEEE Transactions on Cybernetics, 45 (4). pp. 622-634. ISSN 2168-2267 Prototypes Training Reliability Prediction algorithms Cybernetics Manifolds Standards http://ieeexplore.ieee.org/document/6847198/ doi:10.1109/TCYB.2014.2332003 doi:10.1109/TCYB.2014.2332003 |
| spellingShingle | Prototypes Training Reliability Prediction algorithms Cybernetics Manifolds Standards Triguero, Isaac Garcia, Salvador Herrera, Francisco SEG-SSC: a framework based on synthetic examples generation for self-labeled semi-supervised classification |
| title | SEG-SSC: a framework based on synthetic examples generation for self-labeled semi-supervised classification |
| title_full | SEG-SSC: a framework based on synthetic examples generation for self-labeled semi-supervised classification |
| title_fullStr | SEG-SSC: a framework based on synthetic examples generation for self-labeled semi-supervised classification |
| title_full_unstemmed | SEG-SSC: a framework based on synthetic examples generation for self-labeled semi-supervised classification |
| title_short | SEG-SSC: a framework based on synthetic examples generation for self-labeled semi-supervised classification |
| title_sort | seg-ssc: a framework based on synthetic examples generation for self-labeled semi-supervised classification |
| topic | Prototypes Training Reliability Prediction algorithms Cybernetics Manifolds Standards |
| url | https://eprints.nottingham.ac.uk/45410/ https://eprints.nottingham.ac.uk/45410/ https://eprints.nottingham.ac.uk/45410/ |