Adjacent inputs with different labels and hardness in supervised learning

Corresponding author: Sebastián A. Grillo (sebastian.grillo@ua.edu.py)

Saved in:
書目詳細資料
主要作者: Grillo, Sebastián Alberto (author)
其他作者: Mello Román, Julio César (author), Mello Román, Jorge Daniel (author), Vázquez Noguera, José Luis (author), García Torres, Miguel (author), Divina, Federico (author), Gardel Sotomayor, Pedro Esteban (author)
格式: article
語言:英语
出版: 2021
主題:
在線閱讀:https://doi.org/10.1109/ACCESS.2021.3131150
http://hdl.handle.net/20.500.14066/4519
標簽: 添加標簽
沒有標簽, 成為第一個標記此記錄!
_version_ 1870612073263661056
author Grillo, Sebastián Alberto
author2 Mello Román, Julio César
Mello Román, Jorge Daniel
Vázquez Noguera, José Luis
García Torres, Miguel
Divina, Federico
Gardel Sotomayor, Pedro Esteban
author2_role author
author
author
author
author
author
author_browse Divina, Federico
García Torres, Miguel
Gardel Sotomayor, Pedro Esteban
Grillo, Sebastián Alberto
Mello Román, Jorge Daniel
Mello Román, Julio César
Vázquez Noguera, José Luis
author_facet Grillo, Sebastián Alberto
Mello Román, Julio César
Mello Román, Jorge Daniel
Vázquez Noguera, José Luis
García Torres, Miguel
Divina, Federico
Gardel Sotomayor, Pedro Esteban
author_role author
bitstream.checksum.fl_str_mv 91605948b28e0b058521fdd8775aa060
bab9e8031f88f8e0e33175965c583037
0175ea4a2d4caec4bbcc37e300941108
858b22fda432bd774e469302988c1974
cb05490dcbec20bab461d7a20711bab5
bitstream.checksumAlgorithm.fl_str_mv MD5
MD5
MD5
MD5
MD5
bitstream.url.fl_str_mv http://repositorio.conacyt.gov.py/bitstream/20.500.14066/4519/4/Adjacent%20inputs%20with%20different%20labels%20and%20hardness%20in%20supervised%20learning.jpg
http://repositorio.conacyt.gov.py/bitstream/20.500.14066/4519/1/Adjacent%20inputs%20with%20different%20labels%20and%20hardness%20in%20supervised%20learning.pdf
http://repositorio.conacyt.gov.py/bitstream/20.500.14066/4519/2/license_rdf
http://repositorio.conacyt.gov.py/bitstream/20.500.14066/4519/3/license.txt
http://repositorio.conacyt.gov.py/bitstream/20.500.14066/4519/5/Adjacent%20inputs%20with%20different%20labels%20and%20hardness%20in%20supervised%20learning.pdf.txt
dc.contributor.other.es.fl_str_mv Universidad Americana/ INCADE S.A.E
dc.creator.none.fl_str_mv Grillo, Sebastián Alberto
Mello Román, Julio César
Mello Román, Jorge Daniel
Vázquez Noguera, José Luis
García Torres, Miguel
Divina, Federico
Gardel Sotomayor, Pedro Esteban
dc.date.accessioned.none.fl_str_mv 2025-02-03T17:14:44Z
dc.date.available.none.fl_str_mv 2025-02-03T17:14:44Z
dc.date.issued.none.fl_str_mv 2021-11-25
dc.identifier.citation.en.fl_str_mv Grillo, S. A., Mello Román, J. C., Mello Román, J. D., Vázquez Noguera, J. L., García Torres, M., & Divina, F. (2021). Adjacent inputs with different labels and hardness in supervised learning. IEEE Access, 9, 162487-162498. https://doi.org/10.1109/ACCESS.2021.3131150
dc.identifier.doi.es.fl_str_mv 10.1109/ACCESS.2021.3131150
dc.identifier.essn.es.fl_str_mv 2169-3536
dc.identifier.other.es.fl_str_mv https://doi.org/10.1109/ACCESS.2021.3131150
dc.identifier.uri.none.fl_str_mv http://hdl.handle.net/20.500.14066/4519
dc.language.iso.es.fl_str_mv eng
dc.publisher.es.fl_str_mv Institute of Electrical and Electronics Engineers
dc.relation.projectCONACYT.es.fl_str_mv PINV18-1199
dc.rights.*.fl_str_mv Atribución 4.0 Internacional
dc.rights.accessRights.es.fl_str_mv info:eu-repo/semantics/openAccess
dc.rights.copyright.es.fl_str_mv © 2021 Los Autores
dc.rights.uri.*.fl_str_mv http://creativecommons.org/licenses/by/4.0/
dc.subject.classification.es.fl_str_mv 7. Salud
7.3. Prevención, vigilancia y control de enfermedades transmisibles y no transmisibles
dc.subject.en.fl_str_mv Complexity theory
Data models
Machine learning algorithms
Measurement uncertainty
Neural networks
Particle measurements
Supervised learning
dc.subject.ocde.es.fl_str_mv 2. Ingeniería y Tecnología
2.2. Ingeniería Eléctrica, Electrónica e Informática [ingeniería eléctrica, electrónica, ingeniería y sistemas de comunicación, ingeniería informática (sólo equipos) y otras disciplinas afines]
dc.subject.other.es.fl_str_mv Classification
Data complexity
Machine learning
Overfitting
Supervised learning
dc.title.es.fl_str_mv Adjacent inputs with different labels and hardness in supervised learning
dc.type.es.fl_str_mv info:eu-repo/semantics/article
info:eu-repo/semantics/publishedVersion
description Corresponding author: Sebastián A. Grillo (sebastian.grillo@ua.edu.py)
eu_rights_str_mv openAccess
format article
id CONACYT_f070d1cf7f5e5af3cbbc027119fc7769
identifier_str_mv Grillo, S. A., Mello Román, J. C., Mello Román, J. D., Vázquez Noguera, J. L., García Torres, M., & Divina, F. (2021). Adjacent inputs with different labels and hardness in supervised learning. IEEE Access, 9, 162487-162498. https://doi.org/10.1109/ACCESS.2021.3131150
10.1109/ACCESS.2021.3131150
2169-3536
language eng
network_acronym_str CONACYT
network_name_str Repositorio Institucional CONACYT
oai_identifier_str oai:repositorio.conacyt.gov.py:20.500.14066/4519
publishDate 2021
publishDateSort 2021
repository.mail.fl_str_mv repositorio.institucional@conacyt.gov.py
repository.name.fl_str_mv Repositorio Institucional CONACYT
repository_id_str
rights_invalid_str_mv Atribución 4.0 Internacional
http://creativecommons.org/licenses/by/4.0/
© 2021 Los Autores
spelling 57160011bddad9-f33b-43c9-8169-2cb6feaa29de60040eed708-3fa3-4ba3-b01d-29ebaa60cdb1600114460001af5a35-e2ad-4be2-8211-1ee4e7ab15cd60032a48873-af97-439e-bf26-eb8c2883849b600477c754f-d9b7-4bd1-a9f0-7cc92aec11f3600Universidad Americana/ INCADE S.A.E2025-02-03T17:14:44Z2025-02-03T17:14:44Z2021-11-25Grillo, S. A., Mello Román, J. C., Mello Román, J. D., Vázquez Noguera, J. L., García Torres, M., & Divina, F. (2021). Adjacent inputs with different labels and hardness in supervised learning. IEEE Access, 9, 162487-162498. https://doi.org/10.1109/ACCESS.2021.3131150https://doi.org/10.1109/ACCESS.2021.3131150http://hdl.handle.net/20.500.14066/451910.1109/ACCESS.2021.31311502169-3536Corresponding author: Sebastián A. Grillo (sebastian.grillo@ua.edu.py)An important aspect of the design of effective machine learning algorithms is the complexity analysis of classification problems. In this paper, we propose a study aimed at determining the relation between the number of adjacent inputs with different labels and the required number of examples for the task of inducing a classification model. To this aim, we first quantified the adjacent inputs with different labels as a property, using a measure denoted as Neighbour Input Variation (NIV). We analyzed the relation that NIV has to random data and overfitting. We then demonstrated that a threshold of NIV may determine if a classification model can generalize to unseen data. We also presented a case study aimed at analyzing threshold neural networks and the required first hidden layer size in function of NIV. Finally, we performed experiments with five popular algorithms analyzing the relation between NIV and the classification error on problems with few dimensions. We conclude that functions whose similar inputs have different outputs with high probability, considerably reduce the generalization capacity of classification algorithms.Consejo Nacional de Ciencia y TecnologíaPrograma Paraguayo para el Desarrollo de la Ciencia y Tecnología. Proyectos de investigación y desarrolloengInstitute of Electrical and Electronics EngineersAtribución 4.0 Internacionalhttp://creativecommons.org/licenses/by/4.0/info:eu-repo/semantics/openAccess© 2021 Los AutoresComplexity theoryData modelsMachine learning algorithmsMeasurement uncertaintyNeural networksParticle measurementsSupervised learning7. Salud7.3. Prevención, vigilancia y control de enfermedades transmisibles y no transmisiblesClassificationData complexityMachine learningOverfittingSupervised learning2. Ingeniería y Tecnología2.2. Ingeniería Eléctrica, Electrónica e Informática [ingeniería eléctrica, electrónica, ingeniería y sistemas de comunicación, ingeniería informática (sólo equipos) y otras disciplinas afines]Adjacent inputs with different labels and hardness in supervised learninginfo:eu-repo/semantics/articleinfo:eu-repo/semantics/publishedVersionIEEE Access162487162498PINV18-11999Grillo, Sebastián AlbertoMello Román, Julio CésarMello Román, Jorge DanielVázquez Noguera, José LuisGarcía Torres, MiguelDivina, FedericoGardel Sotomayor, Pedro EstebanTHUMBNAILAdjacent inputs with different labels and hardness in supervised learning.jpgAdjacent inputs with different labels and hardness in supervised learning.jpgVista de primera página de artículo científicoimage/jpeg1109822http://repositorio.conacyt.gov.py/bitstream/20.500.14066/4519/4/Adjacent%20inputs%20with%20different%20labels%20and%20hardness%20in%20supervised%20learning.jpg91605948b28e0b058521fdd8775aa060MD54ORIGINALAdjacent inputs with different labels and hardness in supervised learning.pdfAdjacent inputs with different labels and hardness in supervised learning.pdfArtículo científicoapplication/pdf1894441http://repositorio.conacyt.gov.py/bitstream/20.500.14066/4519/1/Adjacent%20inputs%20with%20different%20labels%20and%20hardness%20in%20supervised%20learning.pdfbab9e8031f88f8e0e33175965c583037MD51CC-LICENSElicense_rdflicense_rdfapplication/rdf+xml; charset=utf-8908http://repositorio.conacyt.gov.py/bitstream/20.500.14066/4519/2/license_rdf0175ea4a2d4caec4bbcc37e300941108MD52LICENSElicense.txtlicense.txttext/plain; charset=utf-81698http://repositorio.conacyt.gov.py/bitstream/20.500.14066/4519/3/license.txt858b22fda432bd774e469302988c1974MD53TEXTAdjacent inputs with different labels and hardness in supervised learning.pdf.txtAdjacent inputs with different labels and hardness in supervised learning.pdf.txtExtracted texttext/plain49224http://repositorio.conacyt.gov.py/bitstream/20.500.14066/4519/5/Adjacent%20inputs%20with%20different%20labels%20and%20hardness%20in%20supervised%20learning.pdf.txtcb05490dcbec20bab461d7a20711bab5MD5520.500.14066/4519oai:repositorio.conacyt.gov.py:20.500.14066/45192026-02-12 19:30:39.493Repositorio Institucional CONACYTrepositorio.institucional@conacyt.gov.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
spellingShingle Adjacent inputs with different labels and hardness in supervised learning
Grillo, Sebastián Alberto
Complexity theory
Data models
Machine learning algorithms
Measurement uncertainty
Neural networks
Particle measurements
Supervised learning
7. Salud
7.3. Prevención, vigilancia y control de enfermedades transmisibles y no transmisibles
Classification
Data complexity
Machine learning
Overfitting
Supervised learning
2. Ingeniería y Tecnología
2.2. Ingeniería Eléctrica, Electrónica e Informática [ingeniería eléctrica, electrónica, ingeniería y sistemas de comunicación, ingeniería informática (sólo equipos) y otras disciplinas afines]
status_str publishedVersion
title Adjacent inputs with different labels and hardness in supervised learning
title_full Adjacent inputs with different labels and hardness in supervised learning
title_fullStr Adjacent inputs with different labels and hardness in supervised learning
title_full_unstemmed Adjacent inputs with different labels and hardness in supervised learning
title_short Adjacent inputs with different labels and hardness in supervised learning
title_sort Adjacent inputs with different labels and hardness in supervised learning
topic Complexity theory
Data models
Machine learning algorithms
Measurement uncertainty
Neural networks
Particle measurements
Supervised learning
7. Salud
7.3. Prevención, vigilancia y control de enfermedades transmisibles y no transmisibles
Classification
Data complexity
Machine learning
Overfitting
Supervised learning
2. Ingeniería y Tecnología
2.2. Ingeniería Eléctrica, Electrónica e Informática [ingeniería eléctrica, electrónica, ingeniería y sistemas de comunicación, ingeniería informática (sólo equipos) y otras disciplinas afines]
url https://doi.org/10.1109/ACCESS.2021.3131150
http://hdl.handle.net/20.500.14066/4519