Robust Text Classification Using a Hysteresis-Driven Extended SRN

Garen Arevian; Christo Panchev

doi:10.1007/978-3-540-74695-9_44

Robust Text Classification Using a Hysteresis-Driven Extended SRN

Garen Arevian, Christo Panchev

Research output: Chapter in Book/Report/Conference proceeding › Conference proceeding

1 Citation (Scopus)

Abstract

Recurrent Neural Network (RNN) models have been shown to perform well on artificial grammars for sequential classification tasks over long-term time-dependencies. However, there is a distinct lack of the application of RNNs to real-world text classification tasks. This paper presents results on the capabilities of extended two-context layer SRN models (xRNN) applied to the classification of the Reuters-21578 corpus. The results show that the introduction of high levels of noise to sequences of words in titles, where noise is defined as the unimportant stopwords found in natural language text, is very robustly handled by the classifiers which maintain consistent levels of performance. Comparisons are made with SRN and MLP models, as well as other existing classifiers for the text classification task.

Original language	English
Title of host publication	Proceedings of the 17th International Conference on Artificial Neural Networks (ICANN)
Place of Publication	Berlin
Publisher	Springer Verlag
Pages	425–434
Number of pages	10
ISBN (Print)	978-3-540-74693-5
DOIs	https://doi.org/10.1007/978-3-540-74695-9_44
Publication status	Published - 2007
Externally published	Yes
Event	17th International Conference on Artificial Neural Networks - Porto, Portugal Duration: 9 Sep 2007 → 13 Sep 2007 Conference number: 17

Conference

Conference	17th International Conference on Artificial Neural Networks
Abbreviated title	ICANN 2007
Country	Portugal
City	Porto
Period	9/09/07 → 13/09/07

Access to Document

10.1007/978-3-540-74695-9_44

Fingerprint Dive into the research topics of 'Robust Text Classification Using a Hysteresis-Driven Extended SRN'. Together they form a unique fingerprint.

Cite this

@inproceedings{9c76501940954644871b3251cdb83f5f,

title = "Robust Text Classification Using a Hysteresis-Driven Extended SRN",

abstract = "Recurrent Neural Network (RNN) models have been shown to perform well on artificial grammars for sequential classification tasks over long-term time-dependencies. However, there is a distinct lack of the application of RNNs to real-world text classification tasks. This paper presents results on the capabilities of extended two-context layer SRN models (xRNN) applied to the classification of the Reuters-21578 corpus. The results show that the introduction of high levels of noise to sequences of words in titles, where noise is defined as the unimportant stopwords found in natural language text, is very robustly handled by the classifiers which maintain consistent levels of performance. Comparisons are made with SRN and MLP models, as well as other existing classifiers for the text classification task.",

author = "Garen Arevian and Christo Panchev",

year = "2007",

doi = "10.1007/978-3-540-74695-9_44",

language = "English",

isbn = "978-3-540-74693-5",

pages = "425–434",

booktitle = "Proceedings of the 17th International Conference on Artificial Neural Networks (ICANN)",

publisher = "Springer Verlag",

address = "Austria",

note = "17th International Conference on Artificial Neural Networks, ICANN 2007 ; Conference date: 09-09-2007 Through 13-09-2007",

}

TY - GEN

T1 - Robust Text Classification Using a Hysteresis-Driven Extended SRN

AU - Arevian, Garen

AU - Panchev, Christo

N1 - Conference code: 17

PY - 2007

Y1 - 2007

N2 - Recurrent Neural Network (RNN) models have been shown to perform well on artificial grammars for sequential classification tasks over long-term time-dependencies. However, there is a distinct lack of the application of RNNs to real-world text classification tasks. This paper presents results on the capabilities of extended two-context layer SRN models (xRNN) applied to the classification of the Reuters-21578 corpus. The results show that the introduction of high levels of noise to sequences of words in titles, where noise is defined as the unimportant stopwords found in natural language text, is very robustly handled by the classifiers which maintain consistent levels of performance. Comparisons are made with SRN and MLP models, as well as other existing classifiers for the text classification task.

AB - Recurrent Neural Network (RNN) models have been shown to perform well on artificial grammars for sequential classification tasks over long-term time-dependencies. However, there is a distinct lack of the application of RNNs to real-world text classification tasks. This paper presents results on the capabilities of extended two-context layer SRN models (xRNN) applied to the classification of the Reuters-21578 corpus. The results show that the introduction of high levels of noise to sequences of words in titles, where noise is defined as the unimportant stopwords found in natural language text, is very robustly handled by the classifiers which maintain consistent levels of performance. Comparisons are made with SRN and MLP models, as well as other existing classifiers for the text classification task.

U2 - 10.1007/978-3-540-74695-9_44

DO - 10.1007/978-3-540-74695-9_44

M3 - Conference proceeding

SN - 978-3-540-74693-5

SP - 425

EP - 434

BT - Proceedings of the 17th International Conference on Artificial Neural Networks (ICANN)

PB - Springer Verlag

CY - Berlin

T2 - 17th International Conference on Artificial Neural Networks

Y2 - 9 September 2007 through 13 September 2007

ER -