Mining Opinion Features in Customer Reviews

By Minqing Hu, B. Liu. AAAI Conference on Artificial Intelligence, 2004

1,487 citations recorded by the source corpus.

View Paper: https://www.semanticscholar.org/paper/64b8553bd62997d332910078bcd5ee74a43f9350

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

1. B. Liu
2. Erik Cambria
3. Bing Liu
4. Soujanya Poria
5. Mohit Bansal
6. Zhe Gan
7. Yoshua Bengio
8. Lei Zhang

Citations in the corpus

• Bug report, feature request, or simply praise? On automatically classifying app reviews (2015, background): “Opinion mining is popular in other domains to analyze opinions about movies, cameras, or blogs [17], [24], [28].”
• A machine learning approach to sentiment analysis in multilingual Web texts (2009, methodology): “An opinion word lexicon was constructed by Riloff et al. (2003), while Hu and Liu (2004) identify words describing the features of a product (e.g., “The camera takes incredible pictures .”) and use them in the classification.”
• Opinion Extraction, Summarization and Tracking in News and Blog Corpora (2006, background): “Hu and Liu (2004) proposed an opinion summarization of products, categorized by the opinion polarity.”
• Modeling online reviews with multi-grain topic models (2008, methodology): “Hu and Liu de.ned an aspect as simply a string and there was no attempt to cluster or infer aspects that are mentioned implicitly, e.g., The amount of stains in the room was overwhelming is about the cleanliness aspect for hotels.”
• Great Explanations: Opinionated Explanations for Recommendations (2015, background): “Separately, singlenoun features are validated by eliminating nouns that are rarely associated with sentiment words in reviews as per [19], since such nouns are unlikely to refer to product features; these extracted features are the review features.”
• Detection of Harassment on Web 2.0 (2009, background): “Extensive work [5, 4, 7, 6] has been done in mining user opinions on products.”
• Entrepreneurial Finance and Moral Hazard: Evidence from Token Offerings (2020, methodology): “Each word in each sentence is first classified as being neutral (xi), positive ( + xi ), negative (xi ), amplifiers (xi), or negators (xi) according to the Hu and Liu (2004) dictionary.”
• SemEval-2014 Task 4: Aspect Based Sentiment Analysis (2014, background): “ABSA is critical in mining and summarizing opinions from on-line reviews (Gamon et al., 2005; Titov and McDonald, 2008; Hu and Liu, 2004a; Popescu and Etzioni, 2005).”
• Estimating Aggregate Consumer Preferences from Online Product Reviews (2010, unknown): “Hu and Liu (2004) apply an association rule approach to identify frequent n-grams in product reviews, which are then used as potential attributes.”
• Latent aspect rating analysis without aspect keyword supervision (2011, background): “However, the aspect information is not always available, and it may be di.cult to pre-de.ne appropriate aspects without a good knowledge about what aspects are actually commented on in the reviews.”
• Survey on mining subjective data on the web (2011, methodology): “4.2 Development of opinion aggregation A typical method for opinion aggregation was proposed by Hu and Liu (2004a).”
• How Do Users Like This Feature? A Fine Grained Sentiment Analysis of App Reviews (2014, background): “such as, movies, cameras and desktop software [7], [22] and blogs [18].”
• SenticNet: A Publicly Available Semantic Resource for Opinion Mining (2010, methodology): “…and Hoffmann 2005)(Somasundaran, Wiebe, and Ruppenhofer 2008), and statistical methods, which consist in calculating the valence of keywords, punc-tuation and word co-occurrence frequencies on the base of a large training corpus (Hu and Liu 2004)(Pang and Lee 2005)(Abbasi, Chen, and Salem 2008).”
• What makes a great MOOC? An interdisciplinary analysis of student retention in online courses (2013, methodology): “For example, Hu and Liu (2004a; 2004b) summarized all the customer reviews of a product employing association rule mining to find frequent ngrams to be used as candidate features and Lee and Bradlow (2007) presented an automatic procedure for obtaining conjoint attributes and levels that list the…”