Semantic aspect discovery for online reviews

Md Hijbul Alam; Sang Keun Lee

doi:10.1109/icdm.2012.65

Semantic aspect discovery for online reviews

Md Hijbul Alam, Sang Keun Lee

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

11 Citations (Scopus)

Abstract

The number of opinions and reviews about different products and services is growing online. Users frequently look for important aspects of a product or service in the reviews. Usually, they are interested in semantic (i.e., sentiment-oriented) aspects. However, extracting semantic aspects with supervised methods is very expensive. We propose a domain independent unsupervised model to extract semantic aspects, and conduct qualitative and quantitative experiments to evaluate the extracted aspects. The experiments show that our model effectively extracts semantic aspects with correlated top words. In addition, the conducted evaluation on aspect sentiment classification shows that our model outperforms other models by 5-7% in terms of macro-average F1.

Original language	English
Title of host publication	Proceedings - 12th IEEE International Conference on Data Mining, ICDM 2012
Pages	816-821
Number of pages	6
DOIs	https://doi.org/10.1109/icdm.2012.65
Publication status	Published - 2012
Publication type	A4 Article in conference proceedings
Event	12th IEEE International Conference on Data Mining, ICDM 2012 - Duration: 1 Jan 2012 → …

Conference

Conference	12th IEEE International Conference on Data Mining, ICDM 2012
Period	1/01/12 → …

Keywords

Aspect discovery
Opinion mining
Sentiment analysis
Topic model

Publication forum classification

Publication forum level 1

Access to Document

10.1109/icdm.2012.65

Cite this

@inproceedings{b85cb643d6654393af33d5362e4af971,

title = "Semantic aspect discovery for online reviews",

abstract = "The number of opinions and reviews about different products and services is growing online. Users frequently look for important aspects of a product or service in the reviews. Usually, they are interested in semantic (i.e., sentiment-oriented) aspects. However, extracting semantic aspects with supervised methods is very expensive. We propose a domain independent unsupervised model to extract semantic aspects, and conduct qualitative and quantitative experiments to evaluate the extracted aspects. The experiments show that our model effectively extracts semantic aspects with correlated top words. In addition, the conducted evaluation on aspect sentiment classification shows that our model outperforms other models by 5-7% in terms of macro-average F1.",

keywords = "Aspect discovery, Opinion mining, Sentiment analysis, Topic model, Aspect discovery, Opinion mining, Sentiment analysis, Topic model",

author = "Alam, {Md Hijbul} and Lee, {Sang Keun}",

year = "2012",

doi = "10.1109/icdm.2012.65",

language = "English",

isbn = "978-0-7695-4905-7",

pages = "816--821",

booktitle = "Proceedings - 12th IEEE International Conference on Data Mining, ICDM 2012",

note = "12th IEEE International Conference on Data Mining, ICDM 2012 ; Conference date: 01-01-2012",

}

TY - GEN

T1 - Semantic aspect discovery for online reviews

AU - Alam, Md Hijbul

AU - Lee, Sang Keun

PY - 2012

Y1 - 2012

N2 - The number of opinions and reviews about different products and services is growing online. Users frequently look for important aspects of a product or service in the reviews. Usually, they are interested in semantic (i.e., sentiment-oriented) aspects. However, extracting semantic aspects with supervised methods is very expensive. We propose a domain independent unsupervised model to extract semantic aspects, and conduct qualitative and quantitative experiments to evaluate the extracted aspects. The experiments show that our model effectively extracts semantic aspects with correlated top words. In addition, the conducted evaluation on aspect sentiment classification shows that our model outperforms other models by 5-7% in terms of macro-average F1.

AB - The number of opinions and reviews about different products and services is growing online. Users frequently look for important aspects of a product or service in the reviews. Usually, they are interested in semantic (i.e., sentiment-oriented) aspects. However, extracting semantic aspects with supervised methods is very expensive. We propose a domain independent unsupervised model to extract semantic aspects, and conduct qualitative and quantitative experiments to evaluate the extracted aspects. The experiments show that our model effectively extracts semantic aspects with correlated top words. In addition, the conducted evaluation on aspect sentiment classification shows that our model outperforms other models by 5-7% in terms of macro-average F1.

KW - Aspect discovery

KW - Opinion mining

KW - Sentiment analysis

KW - Topic model

KW - Aspect discovery

KW - Opinion mining

KW - Sentiment analysis

KW - Topic model

U2 - 10.1109/icdm.2012.65

DO - 10.1109/icdm.2012.65

M3 - Conference contribution

AN - SCOPUS:84874104662

SN - 978-0-7695-4905-7

SP - 816

EP - 821

BT - Proceedings - 12th IEEE International Conference on Data Mining, ICDM 2012

T2 - 12th IEEE International Conference on Data Mining, ICDM 2012

Y2 - 1 January 2012

ER -

Semantic aspect discovery for online reviews

Abstract

Conference

Keywords

Publication forum classification

Access to Document

Fingerprint

Cite this