Introducing Weighted Kernel Classifiers for Handling Imbalanced Paralinguistic Corpora: Snoring, Addressee and Cold

Basit Öğe Kaydı
dc.authorid0000-0003-3424-652X
dc.authorid0000-0001-7947-5508
dc.authorscopusid36241785000
dc.authorscopusid57219469958
dc.authorwosidKarpov, Alexey A/A-8905-2012
dc.contributor.authorKaya, Heysem
dc.contributor.authorKarpov, Alexey A.
dc.date.accessioned2022-05-11T14:15:51Z
dc.date.available2022-05-11T14:15:51Z
dc.date.issued2017
dc.departmentFakülteler, Çorlu Mühendislik Fakültesi, Bilgisayar Mühendisliği Bölümü
dc.description18th Annual Conference of the International-Speech-Communication-Association (INTERSPEECH 2017) -- AUG 20-24, 2017 -- Stockholm, SWEDEN
dc.description.abstractThe field of paralinguistics is growing rapidly with a wide range of applications that go beyond recognition of emotions, laughter and personality. The research flourishes in multiple directions such as signal representation and classification, addressing the issues of the domain. Apart from the noise robustness, an important issue with real life data is the imbalanced nature: some classes of states/traits are under-represented. Combined with the high dimensionality of the feature vectors used in the state-of-the-art analysis systems, this issue poses the threat of over-fitting. While the kernel trick can be employed to handle the dimensionality issue, regular classifiers inherently aim to minimize the misclassification error and hence are biased towards the majority class. A solution to this problem is over sampling of the minority class(es). However, this brings increased memory/computational costs, while not bringing any new information to the classifier. In this work, we propose a new weighting scheme on instances of the original dataset, employing Weighted Kernel Extreme Learning Machine, and inspired from that, introducing the Weighted Partial Least Squares Regression based classifier. The proposed methods are applied on all three 1NTERSPEECH ComParF, 2017 challenge corpora, giving better or competitive results compared to the challenge baselines.
dc.description.sponsorshipInt Speech Commun Assoc, Stockholm Univ, KTH Royal Inst Technol, Karolinska Inst, Amazon Alexa, DiDi, Furhat Robot, Microsoft, EZ Alibaba Grp, CIRRUS LOGIC, CVTE, Google, Baidu, IBM Res, YAHOO Japan, Nuance, Voice Provider, ASM Solut Ltd, Mitsubishi Elect Res Lab, Yandex
dc.description.sponsorshipRFBRRussian Foundation for Basic Research (RFBR) [16-37-60100]; Government of Russia [074-U01]; grant of the President of Russia [MD-254.2017.8]
dc.description.sponsorshipThis work is partially supported by RFBR (project No 16-37-60100), grant of the President of Russia (No. MD-254.2017.8) and by the Government of Russia (grant No 074-U01).
dc.identifier.doi10.21437/Interspeech.2017-653
dc.identifier.endpage3531
dc.identifier.isbn978-1-5108-4876-4
dc.identifier.issn2308-457X
dc.identifier.scopus2-s2.0-85039147052
dc.identifier.scopusqualityN/A
dc.identifier.startpage3527
dc.identifier.urihttps://doi.org/10.21437/Interspeech.2017-653
dc.identifier.urihttps://hdl.handle.net/20.500.11776/6094
dc.identifier.wosWOS:000457505000734
dc.identifier.wosqualityN/A
dc.indekslendigikaynakWeb of Science
dc.indekslendigikaynakScopus
dc.institutionauthorKaya, Heysem
dc.language.isoen
dc.publisherIsca-Int Speech Communication Assoc
dc.relation.ispartof18th Annual Conference of the International Speech Communication Association (Interspeech 2017), Vols 1-6: Situated Interaction
dc.relation.publicationcategoryKonferans Öğesi - Uluslararası - Kurum Öğretim Elemanıen_US
dc.rightsinfo:eu-repo/semantics/closedAccess
dc.subjectcomputational paralinguistics
dc.subjectimbalanced data
dc.subjectSnoring
dc.subjectAddressee
dc.subjectFisher vector
dc.subjectWeighted PLS
dc.subjectELM
dc.subjectExtreme Learning-Machine
dc.subjectDeception
dc.titleIntroducing Weighted Kernel Classifiers for Handling Imbalanced Paralinguistic Corpora: Snoring, Addressee and Cold
dc.typeConference Object

Dosyalar

Orijinal paket
Listeleniyor 1 - 1 / 1
Küçük Resim Yok
İsim:
6094.pdf
Boyut:
122.27 KB
Biçim:
Adobe Portable Document Format
Açıklama:
Tam Metin / Full Text
İndir

Koleksiyon

WoS İndeksli Yayınlar Koleksiyonu
Scopus İndeksli Yayınlar Koleksiyonu
Çorlu Mühendislik Fakültesi Koleksiyonu