Current Visitors:


Product feature grouping for opinion mining

Zhongwu Zhai, Bing Liu, Jingyuan Wang, Hua Xu and Peifa Jia

IEEE Intelligent Systems, vol. 27, no. 4, pp. 37–44, 2011. Download

Review aggregators and e-commerce sites are just two examples of businesses that rely on opinion mining to produce feature-based summaries of products' qualities.1 This model first identifies product features and then collects positive and negative opinions on them to produce a summary of good and bad points. Features of a product are attributes, components, and other aspects, such as “picture quality” and “zoom” for a digital camera.

Reviewers often use different words or phrases to describe the same product feature. For example, “picture” and “photo” mean the same thing for cameras. Grouping such synonyms is critical for effective opinion summary. Although WordNet and other thesauri can help to some extent, they aren't enough. For one thing, many words and phrases that are not synonyms in a dictionary may refer to the same feature in an application domain-for example, “appearance” and “design” are not synonymous, but they might both describe how something looks. For another, many synonyms are domain-dependent: “movie” and “picture” are synonyms in movie reviews but not in camera reviews, where “movie” is more likely to mean “video.” And finally, determining which expressions indicate the same feature can depend on the user's application need. For example, in car reviews, internal design and external design can be two separate features or just one, called “design,” according to the level of detail the user needs. In camera reviews, one shopper may want to study the battery as a whole, while another wants to know about battery weight, and battery life separately. For this reason, in applications the user needs to be involved in synonym grouping.

Influence of the context window size on SC-EM
Influence of the context window size on SC-EM

If you find our work is helpful for your research, please kindly consider citing our paper.

 

@article{zhai2011product,

  title={Product feature grouping for opinion mining},

  author={Zhai, Zhongwu and Liu, Bing and Wang, Jingyuan and Xu, Hua and Jia, Peifa},

  journal={IEEE Intelligent Systems},

  volume={27},

  number={4},

  pages={37--44},

  year={2011},

  publisher={IEEE}

}