Academic
Publications
Mining Opinion Features in Customer Reviews

Mining Opinion Features in Customer Reviews,Minqing Hu,Bing Liu

Mining Opinion Features in Customer Reviews   (Citations: 174)
BibTex | RIS | RefWorks Download
It is a common practice that merchants selling products on the Web ask their customers to review the products and associated services. As e-commerce is becoming more and more popular, the number of customer reviews that a product receives grows rapidly. For a popular product, the number of reviews can be in hundreds. This makes it difficult for a potential customer to read them in order to make a decision on whether to buy the product. In this project, we aim to summarize all the customer reviews of a product. This summarization task is different from traditional text summarization because we are only interested in the specific features of the product that customers have opinions on and also whether the opinions are positive or negative. We do not summarize the reviews by selecting or rewriting a subset of the original sentences from the reviews to capture their main points as in the classic text summarization. In this paper, we only focus on mining opinion/product features that the reviewers have commented on. A number of techniques are presented to mine such features. Our experimental results show that these techniques are highly effective.
Cumulative Annual
View Publication
The following links allow you to view full publications. These links are maintained by other sources not affiliated with Microsoft Academic Search.
    • ...Hu and Liu’s work [14, 15] as well as our previous work [22] considered adjectives to be opinions...
    • ...While Hu and Liu [15] also made use of POS tags for this task, they adapted the idea of frequent...
    • ...The most commonly-used SO determination approach operates on the assumption that semantic similarity between words implies their sentimental similarity .H u and Liu [14, 15] presented a bootstrapping algorithm that makes use of a seed set containing opinion-bearing words having known SO, such as “great” for Positive and “bad” for Negative...

    Cane Wing-ki Leunget al. A probabilistic rating inference framework for mining user preferences...

    • ...They use the association rule mining approach presented in [14] to obtain a first list of features...

    Jorge Carrillo-de-Albornozet al. A Joint Model of Feature Mining and Sentiment Analysis for Product Rev...

    • ...Considering that some features usually delivered positive evaluations on the product while others did the contrary, Hu and Liu proposed a method aiming at mining relationships between user opinions and product features [8]...

    Wei Zhenget al. Music Review Classification Enhanced by Semantic Information

    • ...Two main reasons are that the terminology used by manufactures may be different to the terms used by customers and customers may comment on unexpected features that manufactures have never thought about [5]...
    • ...To deal the task of production feature extraction, the authors of [5] generate a set of frequent features by finding out frequent terms as preliminary feature set then prune the feature set by calculating compactness and redundancy of term phrases...
    • ...Major applications of opinion mining related techniques are product review mining [5,9,13,22], recommendation systems [23] and business intelligence [2]...
    • ...One approach is to develop linguistic resources of sentiment orientation and structures of sentiment expression as in [5]...
    • ...One is frequent noun/phrases proposed in [5]...

    Yung-Ming Liet al. Deriving Marketing Intelligence over Microblogs

    • ...Many recent studies on opinion analysis aimed to analyze and extract opinions or sentiment information from customer reviews and present them in the form of sentiment-based or opinion-oriented summarization [16], [17], [28], [44], [21]...
    • ...One representative work of such techniques is featurebased sentiment summarization [16], [17]...
    • ...Third, Hu and Liu [16], [17] used stemming, fuzzy matching, and WordNetbased synonym finding techniques to deal with the problem of word variants and misspellings for product feature generalization...
    • ...Hu and Liu ’s [16], [17] techniques cannot effectively tackle this problem...
    • ...Recent work of aspect-based opinion analysis has generally focused on product reviews [28], [18], [13], [22], [29], [16], [17], [8]...
    • ...The most related work is feature-based sentiment summarization [16], [17], which aimed at producing a summary expressing the aggregated sentiment for each feature of a product and supporting textual evidence...
    • ...Hu and Liu [16], [17] only focused on mining features that explicitly appear as nouns or noun phrases in the product reviews and cannot effectively deal with the implicit feature expression problem...
    • ...Second, Hu and Liu [16], [17] assumed that product features and opinion words explicitly appear as noun phrases and adjectives, respectively...
    • ...Third, Hu and Liu [16], [17] used stemming and fuzzy matching techniques to deal with the problem of word variants and misspellings for feature generalization...
    • ...Unlike our aspect-based sentence segmentation, the method of Hu and Liu [16], [17] identified one or more features expressed by a reviewing sentence, but still used the whole sentence (instead of segments) for feature-based summarization generation...
    • ...In related areas of opinion extraction from user reviews, some previous efforts have focused on the extraction of opinion topics [16], [28] that is limited to extracting the mentions of product names and their features...

    Jingbo Zhuet al. Aspect-Based Opinion Polling from Customer Reviews

Sort by: