Online reviews provide consumers with valuable information that guides their decisions on a variety of fronts: from entertainment and shopping to medical services. Although the proliferation of online reviews gives insights about different aspects of a product, it can also prove a serious drawback: consumers cannot and will not read thousands of reviews before making a purchase decision. This need to extract useful information from large review corpora has spawned considerable prior work, but so far all have drawbacks. Review summarization (generating statistical descriptions of review sets) sacrifices the immediacy and narrative structure of reviews. Likewise, review selection (identifying a subset of 'helpful' or 'important' reviews) leads to redundant or non-representative summaries. In this paper, we fill the gap between existing review-summarization and review-selection methods by selecting a small subset of reviews that together preserve the statistical properties of the entire review corpus. We formalize this task as a combinatorial optimization problem and show that it NP-hard both tosolve and approximate. We also design effective algorithms that prove to work well in practice. Our experiments with real review corpora on different types of products demonstrate the utility of our methods, and our user studies indicate that our methods provide a better summary than prior approaches.

/pdf/selecting-a-characteristic-set-of-reviews-4ffig0q0ri.pdf

Selecting a characteristic set of reviews

Research about some aspect-based opinion summarization methods.A new content selection strategy to produce extractive summaries is proposed.A novel NLG template-based system to generate abstractive summaries is proposed.Extractive and abstractive opinion summarization methods are compared. In the last years, the opinion summarization task has gained much importance because of the large amount of online information and the increasing interest in learning the user evaluation about products, services, companies, and people. Although there are many works in this area, there is room for improvement, as the results are far from ideal. In this paper, we present our investigations to generate extractive and abstractive summaries of opinions. We study some well-known methods in the area and compare them. Besides using these methods, we also develop new methods that consider the main advantages of the ones before. We evaluate them according to three traditional summarization evaluation measures: informativeness, linguistic quality, and utility of the summary. We show that we produce interesting results and that our methods outperform some methods from literature.

Opinion summarization methods

Fusing and mining opinions for reputation generation

Tourism for a local city is one of the most important key industries. The Web contains much information for the tourism, such as impressions and sentiments about sightseeing areas. Analyzing the information is a significant task for tourism informatics. In this paper, we propose a tourism information analysis system for a local city. The target resource for the analysis is information on Twitter. First, we discuss a method to extract tweets (posted sentences) relating to the target locations and tourism events. Then, we analyze the polarity of the extracted tweets; positive or negative opinions. It is well-known as a P/N classification task in natural language processing. For the process, we employ an unsupervised machine learning approach that uses seed words. We evaluate and consider the extraction and P/N classification tasks. The experimental result about P/N classification shows the effectiveness of our method.

/pdf/analyzing-tourism-information-on-twitter-for-a-local-city-txqya7b9bm.pdf

Analyzing Tourism Information on Twitter for a Local City

Product reviews greatly influence purchase decisions in online shopping. A common burden of online shopping is that consumers have to search for the right answers through massive reviews, especially on popular products. Hence, estimating and predicting the helpfulness of reviews become important tasks to directly improve shopping experience. In this paper, we propose a new approach to helpfulness prediction by leveraging aspect analysis of reviews. Our hypothesis is that a helpful review will cover many aspects of a product at different emphasis levels. The first step to tackle this problem is to extract proper aspects. Because related products share common aspects to different degrees, we propose an aspect extraction model making use of product category information to balance the aspects of a general category and those of subcategories under it. On top of this model, a two-layer regressor is trained for helpfulness prediction. Experiment results show that we can improve helpfulness prediction by 7% than the baseline on 5 popular product categories from Amazon.com.

https://ink.library.smu.edu.sg/cgi/viewcontent.cgi?article=7020&context=sis_research

Aspect-Based Helpfulness Prediction for Online Product Reviews

The development of the Web services lets many users easily provide their opinions recently. Automatic summarization of enormous sentiments has been expected. Intuitively, we can summarize a review with traditional document summarization methods. However, such methods have not well-discussed "aspects". Basically, a review consists of sentiments with various aspects. We summarize reviews for each aspect so that the summary presents information without biasing to a specific topic. In this paper, we propose a method for multi- aspects review summarization based on evaluative sentence extraction. We handle three fea- tures; ratings of aspects, the tf -idf value, and the number of mentions with a similar topic. For estimating the number of mentions, we apply a clustering algorithm. By integrating these features, we generate a more appropriate summary. The experiment results show the effectiveness of our method.

/pdf/multi-aspects-review-summarization-based-on-identification-l66j1q16ou.pdf

Multi-aspects Review Summarization Based on Identification of Important Opinions and their Similarity

Multi-aspects review summarization with objective information

As the World Wide Web rapidly grows, a huge number of online documents are easily accessible on the Web. We obtain a huge number of review documents that include user’s opinions for products. To summarize the opinions is one of the hottest topics in natural language processing. We focus on aspects of a product in the summarization process. First, we identify a relation between aspects and each word in review documents. Our method employs an unsupervised approach for the identification process. The experimental result shows the effectiveness of the method. Next, we generate a summary by using the relations. In our system, users obtain the summary by an interactive approach.

http://www.pluto.ai.kyutech.ac.jp/~shimada/paper/SCIS_ISIS2010.pdf

Ryosuke Tadano

Papers

Multi-aspects Review Summarization Based on Identification of Important Opinions and their Similarity

Multi-aspects review summarization with objective information

Interactive Aspect Summarization Using Word-aspect Relations for Review Documents