Extension of conventional co-training learning strategies to three-view and committee-based learning strategies for effective automatic sentence segmentation
Yükleniyor...
Dosyalar
Tarih
2018
Yazarlar
Dergi Başlığı
Dergi ISSN
Cilt Başlığı
Yayıncı
IEEE
Erişim Hakkı
info:eu-repo/semantics/closedAccess
Özet
The objective of this work is to develop effective multi-view semi-supervised machine learning strategies for sentence boundary classification problem when only small sets of sentence boundary labeled data are available. We propose three-view and committee-based learning strategies incorporating with co-training algorithms with agreement, disagreement, and self-combined learning strategies using prosodic, lexical and morphological information. We compare experimental results of proposed three-view and committee-based learning strategies to other semi-supervised learning strategies in the literature namely, self-training and co-training with agreement, disagreement, and self-combined strategies. The experiment results show that sentence segmentation performance can be highly improved using multi-view learning strategies that we propose since data sets can be represented by three redundantly sufficient and disjoint feature sets. We show that the proposed strategies substantially improve the average performance when only a small set of manually labeled data is available for Turkish and English spoken languages, respectively.
Açıklama
Anahtar Kelimeler
Boosting, Co-training, Sentence segmentation, Semi-supervised learning, Prosody, Speech, Learning algorithms, Machine learning, Supervised learning, Data models, Semisupervised learning, Feature extraction, Training, Tools, Task analysis, Learning (artificial intelligence), Natural language processing, Speech processing, Multiview learning strategies, Disjoint feature sets, Manually labeled data, Sentence boundary classification problem, Sentence boundary labeled data, Committee-based learning strategies, Prosodic information, Lexical information, Morphological information, Self-combined strategies, Automatic sentence segmentation, Conventional co-training learning, Multiview semisupervised machine learning, Turkish spoken languages, English spoken languages
Kaynak
2018 IEEE Spoken Language Technology Workshop (SLT)
WoS Q Değeri
N/A
Scopus Q Değeri
N/A
Cilt
Sayı
Künye
Dalva, D., Güz, Ü. & Gürkan, H. (2018). Extension of conventional co-training learning strategies to three-view and committee-based learning strategies for effective automatic sentence segmentation. Paper presented at the 2018 IEEE Spoken Language Technology Workshop (SLT), 750-755. doi:10.1109/SLT.2018.8639533
