With the increase in popularity of online review sites comes a corresponding need for tools capable of extracting the information most important to the user from the plain text data. Due to the diversity in products and services being reviewed, supervised methods are often not practical. We present an unsupervised system for extracting aspects and determining sentiment in review text. The method is simple and flexible with regard to domain and language, and takes into account the influence of aspect on sentiment polarity, an issue largely ignored in previous literature. We demonstrate its effectiveness on both component tasks, where it achieves similar results to more complex semi-supervised methods that are restricted by their reliance on manual annotation and extensive knowledge sources. 1

The problem addressed in this paper is to predict a user’s numeric rating in a product review from the text of the review. Unigram and n-gram representations of text are common choices in opinion mining. However, unigrams cannot capture important expressions like “could have been better”, which are essential for prediction models of ratings. N-grams of words, on the other hand, capture such phrases, but typically occur too sparsely in the training set and thus fail to yield robust predictors. This paper overcomes the limitations of these two models, by introducing a novel kind of bag-of-opinions representation, where an opinion, within a review, consists of three components: a root word, a set of modifier words from the same sentence, and one or more negation words. Each opinion is assigned a numeric score which is learned, by ridge regression, from a large, domain-independent corpus of reviews. For the actual test case of a domain-dependent review, the review’s rating is predicted by aggregating the scores of all opinions in the review and combining it with a domaindependent unigram model. The paper presents a constrained ridge regression algorithm for learning opinion scores. Experiments show that the bag-of-opinions method outperforms prior state-of-the-art techniques for review rating prediction.

In this paper, we propose an efficient embedding for modeling higherorder (n-gram) phrases that projects the n-grams to low-dimensional latent semantic space, where a classification function can be defined. We utilize a deep neural network to build a unified discriminative framework that allows for estimating the parameters of the latent space as well as the classification function with a bias for the target classification task at hand. We apply the framework to large-scale sentimental classification task. We present comparative evaluation of the proposed method on two (large) benchmark data sets for online product reviews. The proposed method achieves superior performance in comparison to the state of the art.

Abstract—The majority of online reviews consist of plaintext feedback together with a single numeric score. However, there are multiple dimensions to products and opinions, and understanding the ‘aspects ’ that contribute to users ’ ratings may help us to better understand their individual preferences. For example, a user’s impression of an audiobook presumably depends on aspects such as the story and the narrator, and knowing their opinions on these aspects may help us to recommend better products. In this paper, we build models for rating systems in which such dimensions are explicit, in the sense that users leave separate ratings for each aspect of a product. By introducing new corpora consisting of five million reviews, rated with between three and six aspects, we evaluate our models on three prediction tasks: First, we use our model to uncover which parts of a review discuss which of the rated aspects. Second, we use our model to summarize reviews, which for us means finding the sentences that best explain a user’s rating. Finally, since aspect ratings are optional in many of the datasets we consider, we use our model to recover those ratings that are missing from a user’s evaluation. Our model matches state-of-the-art approaches on existing small-scale datasets, while scaling to the real-world datasets we introduce. Moreover, our model is able to ‘disentangle ’ content and sentiment words: we automatically learn content words that are indicative of a particular aspect as well as the aspect-specific sentiment words that are indicative of a particular rating. Keywords-machine learning; segmentation; summarization; sentiment analysis I.