Beyond the Stars: Improving Rating Predictions using Review Text Content

By Gayatree Ganu, Noémie Elhadad, Amélie Marian. 2009

492 citations recorded by the source corpus.

How researchers see this paper

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

Citations in the corpus

Why Amazon's Ratings Might Mislead You: The Story of Herding Effects (2014, neutral): “While they are useful for review spam detection, 20 customer sentiment analysis, 21,22 product recommendation, 23 and more, insights extracted from semantic features are, however, not mechanistic, and hence not capable of projecting the full rating trajectories.”
Document Modeling with Gated Recurrent Neural Network for Sentiment Classification (2015, background): “Representative features include word ngrams (Wang and Manning, 2012), text topic (Ganu et al., 2009), bag-of-opinions (Qu et al.”
Learning Attitudes and Attributes from Multi-aspect Reviews (2012, methodology): “We assume that each sentence in a review discusses a single aspect; alternately we could model one aspect per word or per paragraph , though one aspect per sentence matches what appears in existing work [5], [12], [30].”
Learning Semantic Representations of Users and Products for Document Level Sentiment Classification (2015, background): “Many studies design effective features, such as text topic (Ganu et al., 2009), bag-of-opinion (Qu et al.”
MMALFM (2018, background): “An early work [21] in this direction relies on domain knowledge to manually label reviews into different aspects, which requires expensive domain knowledge and high labor cost.”
TopicMF: Simultaneously Exploiting Ratings and Reviews for Recommendation (2014, background): “For example, Ganu et al. (2009) depend on the domain knowledge provided by human annotators to extract “explicit” aspects information (e.”
Jointly Modeling Aspects and Opinions with a MaxEnt-LDA Hybrid (2010, methodology): “Ganu et al. (2009) provide a set of annotated sentences from the restaurant data set, in which each sentence has been assigned one or more labels from a gold standard label set S = {Staff, Food, Ambience, Price, Anecdote, Misc}.”
SemEval-2016 Task 5: Aspect Based Sentiment Analysis (2016, background): “Currently, most of the available SA-related datasets, whether released in the context of shared tasks or not (Socher et al., 2013; Ganu et al., 2009), are monolingual and usually focus on English texts.”
Aspect-Aware Latent Factor Model: Rating Prediction with Ratings and Reviews (2018, background): “An early work [14] in this direction relied on domain knowledge to manually label reviews into different aspects, which requires expensive domain knowledge and high labor cost.”
Inferring Networks of Substitutable and Complementary Products (2015, background): “A variety of works have used topic models within recommender systems ( e.g. [6, 10, 11, 22, 23, 28, 31, 32]), though generally with the goal of predicting user ratings (or opinions) rather than learning relationships between products as we do here.”
SemEval-2015 Task 12: Aspect Based Sentiment Analysis (2015, methodology): “Several ABSA methods have been proposed for various domains, like consumer electronics (Hu and Liu {2004a, 2004b}), restaurants (Ganu et al., 2009) and movies (Thet et al.”
Sentiment Analysis in Tourism: Capitalizing on Big Data (2019, methodology): “It is also noted that for subjectivity detection, SVM, Naïve Bayes and k-nearest neighbour methods with regard to machine learning approach, and dictionary-based approach were applied to tourism (Schmunk et al., 2014).”
SemEval-2014 Task 4: Aspect Based Sentiment Analysis (2014, methodology): “…of this kind have been developed for movie reviews (Thet et al., 2010), customer reviews of electronic products like digital cameras (Hu and Liu, 2004a) or net-book computers (Brody and Elhadad, 2010), services (Long et al., 2010), and restaurants (Ganu et al., 2009; Brody and Elhadad, 2010).”
Yelp Dataset Challenge: Review Rating Prediction (2016, background): “In logistic regression (Freedman, 2009), the conditional probability function P ( s | r ) is modelled, where r is a feature vector for review r and s belongs to the set of class labels { 1 , 2 , 3 , 4 , 5 } .”
Quantifying herding effects in crowd wisdom (2014, background): “Finally, there have been a number of interesting studies into the semantics of collective opinions, such as that analyze the text and social aspects of product reviews [10, 15, 21, 9].”
Rating-Boosted Latent Topics: Understanding Users and Items with Ratings and Reviews (2016, background): “Ganu et al. attempt to extract aspect information (e.g., price) manually depending on abundant domain knowledge [Ganu et al., 2009].”
Ratings meet reviews, a combined approach to recommend (2014, background): “Due to the advantage of taking the review text into consideration, there are a few efforts [2, 4, 16, 17] explored the combination of content-based filtering and collaborative filtering.”
Explicit factor models for explainable recommendation based on phrase-level sentiment analysis (2014, background): “More recent work leverages review- or sentence-level sentiment analysis to boost the performance of rating prediction [12, 13, 7, 27, 26].”
A Challenge Dataset and Effective Models for Aspect-Based Sentiment Analysis (2019, methodology): “Data Collection Similar to SemEval-2014 Restaurant Review dataset (Pontiki et al., 2014 York dataset collected by (Ganu et al., 2009).”
Improving the quality of predictions using textual information in online user reviews (2013, methodology): “In [12], our preliminary results indicated that using text for rating prediction was a promising direction.”