A Dual Knowledge Aggregation Network for Cross-Domain Sentiment Analysis

Pengfei Ji; Dandan Song

doi:10.1109/CVIDLICCEA56201.2022.9825235

A Dual Knowledge Aggregation Network for Cross-Domain Sentiment Analysis

Pengfei Ji, Dandan Song

School of Computer Science and Technology

Beijing Institute of Technology

Research output: Chapter in Book/Report/Conference proceeding › Conference contribution › peer-review

1 Citation (Scopus)

Abstract

Cross-domain sentiment analysis (CDSA) is an essential subtask of sentiment analysis. It aims to utilize rich source domain data to conquer the data-hungry problem on target domain. Most existing approaches depending on deep learning mainly concentrate on common features or pivots. However, few of them consider the effect of external Knowledge Graph (KG). In this paper, we propose a Dual Knowledge Aggregation Network for Cross-Domain Sentiment Analysis (DKAN), which leverages prior knowledge from two external KGs. Specifically, DKAN comprises two main parts. One is extracting sentence representation features. The other aims to introduce external knowledge better. Also, we use SenticNet to avoid noise from KG by selecting top-n words and inserting special tokens in sentences. We also conduct empirical analyses on the effectiveness of our model on the Amazon reviews dataset. DKAN achieves promising performance compared with other methods.

Original language	English
Title of host publication	2022 3rd International Conference on Computer Vision, Image and Deep Learning and International Conference on Computer Engineering and Applications, CVIDL and ICCEA 2022
Publisher	Institute of Electrical and Electronics Engineers Inc.
Pages	844-850
Number of pages	7
ISBN (Electronic)	9781665459112
DOIs	https://doi.org/10.1109/CVIDLICCEA56201.2022.9825235
Publication status	Published - 2022
Event	3rd International Conference on Computer Vision, Image and Deep Learning and International Conference on Computer Engineering and Applications, CVIDL and ICCEA 2022 - Virtual, Changchun, China Duration: 20 May 2022 → 22 May 2022

Publication series

Name	2022 3rd International Conference on Computer Vision, Image and Deep Learning and International Conference on Computer Engineering and Applications, CVIDL and ICCEA 2022

Conference

Conference	3rd International Conference on Computer Vision, Image and Deep Learning and International Conference on Computer Engineering and Applications, CVIDL and ICCEA 2022
Country/Territory	China
City	Virtual, Changchun
Period	20/05/22 → 22/05/22

Keywords

SenticNet
corss-domain
knowledge graph
sentiment analysis

Access to Document

10.1109/CVIDLICCEA56201.2022.9825235

Cite this

Ji, P., & Song, D. (2022). A Dual Knowledge Aggregation Network for Cross-Domain Sentiment Analysis. In 2022 3rd International Conference on Computer Vision, Image and Deep Learning and International Conference on Computer Engineering and Applications, CVIDL and ICCEA 2022 (pp. 844-850). (2022 3rd International Conference on Computer Vision, Image and Deep Learning and International Conference on Computer Engineering and Applications, CVIDL and ICCEA 2022). Institute of Electrical and Electronics Engineers Inc.. https://doi.org/10.1109/CVIDLICCEA56201.2022.9825235

Ji, Pengfei ; Song, Dandan. / A Dual Knowledge Aggregation Network for Cross-Domain Sentiment Analysis. 2022 3rd International Conference on Computer Vision, Image and Deep Learning and International Conference on Computer Engineering and Applications, CVIDL and ICCEA 2022. Institute of Electrical and Electronics Engineers Inc., 2022. pp. 844-850 (2022 3rd International Conference on Computer Vision, Image and Deep Learning and International Conference on Computer Engineering and Applications, CVIDL and ICCEA 2022).

@inproceedings{8ca318a1ec024b1a8d70752ed2167624,

title = "A Dual Knowledge Aggregation Network for Cross-Domain Sentiment Analysis",

abstract = "Cross-domain sentiment analysis (CDSA) is an essential subtask of sentiment analysis. It aims to utilize rich source domain data to conquer the data-hungry problem on target domain. Most existing approaches depending on deep learning mainly concentrate on common features or pivots. However, few of them consider the effect of external Knowledge Graph (KG). In this paper, we propose a Dual Knowledge Aggregation Network for Cross-Domain Sentiment Analysis (DKAN), which leverages prior knowledge from two external KGs. Specifically, DKAN comprises two main parts. One is extracting sentence representation features. The other aims to introduce external knowledge better. Also, we use SenticNet to avoid noise from KG by selecting top-n words and inserting special tokens in sentences. We also conduct empirical analyses on the effectiveness of our model on the Amazon reviews dataset. DKAN achieves promising performance compared with other methods.",

keywords = "SenticNet, corss-domain, knowledge graph, sentiment analysis",

author = "Pengfei Ji and Dandan Song",

note = "Publisher Copyright: {\textcopyright} 2022 IEEE.; 3rd International Conference on Computer Vision, Image and Deep Learning and International Conference on Computer Engineering and Applications, CVIDL and ICCEA 2022 ; Conference date: 20-05-2022 Through 22-05-2022",

year = "2022",

doi = "10.1109/CVIDLICCEA56201.2022.9825235",

language = "English",

series = "2022 3rd International Conference on Computer Vision, Image and Deep Learning and International Conference on Computer Engineering and Applications, CVIDL and ICCEA 2022",

publisher = "Institute of Electrical and Electronics Engineers Inc.",

pages = "844--850",

booktitle = "2022 3rd International Conference on Computer Vision, Image and Deep Learning and International Conference on Computer Engineering and Applications, CVIDL and ICCEA 2022",

address = "United States",

}

Ji, P & Song, D 2022, A Dual Knowledge Aggregation Network for Cross-Domain Sentiment Analysis. in 2022 3rd International Conference on Computer Vision, Image and Deep Learning and International Conference on Computer Engineering and Applications, CVIDL and ICCEA 2022. 2022 3rd International Conference on Computer Vision, Image and Deep Learning and International Conference on Computer Engineering and Applications, CVIDL and ICCEA 2022, Institute of Electrical and Electronics Engineers Inc., pp. 844-850, 3rd International Conference on Computer Vision, Image and Deep Learning and International Conference on Computer Engineering and Applications, CVIDL and ICCEA 2022, Virtual, Changchun, China, 20/05/22. https://doi.org/10.1109/CVIDLICCEA56201.2022.9825235

A Dual Knowledge Aggregation Network for Cross-Domain Sentiment Analysis. / Ji, Pengfei; Song, Dandan.
2022 3rd International Conference on Computer Vision, Image and Deep Learning and International Conference on Computer Engineering and Applications, CVIDL and ICCEA 2022. Institute of Electrical and Electronics Engineers Inc., 2022. p. 844-850 (2022 3rd International Conference on Computer Vision, Image and Deep Learning and International Conference on Computer Engineering and Applications, CVIDL and ICCEA 2022).

Research output: Chapter in Book/Report/Conference proceeding › Conference contribution › peer-review

TY - GEN

T1 - A Dual Knowledge Aggregation Network for Cross-Domain Sentiment Analysis

AU - Ji, Pengfei

AU - Song, Dandan

PY - 2022

Y1 - 2022

N2 - Cross-domain sentiment analysis (CDSA) is an essential subtask of sentiment analysis. It aims to utilize rich source domain data to conquer the data-hungry problem on target domain. Most existing approaches depending on deep learning mainly concentrate on common features or pivots. However, few of them consider the effect of external Knowledge Graph (KG). In this paper, we propose a Dual Knowledge Aggregation Network for Cross-Domain Sentiment Analysis (DKAN), which leverages prior knowledge from two external KGs. Specifically, DKAN comprises two main parts. One is extracting sentence representation features. The other aims to introduce external knowledge better. Also, we use SenticNet to avoid noise from KG by selecting top-n words and inserting special tokens in sentences. We also conduct empirical analyses on the effectiveness of our model on the Amazon reviews dataset. DKAN achieves promising performance compared with other methods.

AB - Cross-domain sentiment analysis (CDSA) is an essential subtask of sentiment analysis. It aims to utilize rich source domain data to conquer the data-hungry problem on target domain. Most existing approaches depending on deep learning mainly concentrate on common features or pivots. However, few of them consider the effect of external Knowledge Graph (KG). In this paper, we propose a Dual Knowledge Aggregation Network for Cross-Domain Sentiment Analysis (DKAN), which leverages prior knowledge from two external KGs. Specifically, DKAN comprises two main parts. One is extracting sentence representation features. The other aims to introduce external knowledge better. Also, we use SenticNet to avoid noise from KG by selecting top-n words and inserting special tokens in sentences. We also conduct empirical analyses on the effectiveness of our model on the Amazon reviews dataset. DKAN achieves promising performance compared with other methods.

KW - SenticNet

KW - corss-domain

KW - knowledge graph

KW - sentiment analysis

UR - http://www.scopus.com/inward/record.url?scp=85135396708&partnerID=8YFLogxK

U2 - 10.1109/CVIDLICCEA56201.2022.9825235

DO - 10.1109/CVIDLICCEA56201.2022.9825235

M3 - Conference contribution

AN - SCOPUS:85135396708

T3 - 2022 3rd International Conference on Computer Vision, Image and Deep Learning and International Conference on Computer Engineering and Applications, CVIDL and ICCEA 2022

SP - 844

EP - 850

BT - 2022 3rd International Conference on Computer Vision, Image and Deep Learning and International Conference on Computer Engineering and Applications, CVIDL and ICCEA 2022

PB - Institute of Electrical and Electronics Engineers Inc.

T2 - 3rd International Conference on Computer Vision, Image and Deep Learning and International Conference on Computer Engineering and Applications, CVIDL and ICCEA 2022

Y2 - 20 May 2022 through 22 May 2022

ER -

Ji P, Song D. A Dual Knowledge Aggregation Network for Cross-Domain Sentiment Analysis. In 2022 3rd International Conference on Computer Vision, Image and Deep Learning and International Conference on Computer Engineering and Applications, CVIDL and ICCEA 2022. Institute of Electrical and Electronics Engineers Inc. 2022. p. 844-850. (2022 3rd International Conference on Computer Vision, Image and Deep Learning and International Conference on Computer Engineering and Applications, CVIDL and ICCEA 2022). doi: 10.1109/CVIDLICCEA56201.2022.9825235

A Dual Knowledge Aggregation Network for Cross-Domain Sentiment Analysis

Abstract

Publication series

Conference

Keywords

Access to Document

Other files and links

Fingerprint

Cite this