ABSTRACT
Summarization of Reviews) .............................................................. 214 9.4.3 Content Recommendation I: Popular Podcasts ................................ 214 9.4.4 Content Recommendation II: Helpful Hotel Reviews ...................... 214 9.4.5 Content Recommendation III: Attractive Photos ............................. 215 9.4.6 Predicting Ratings of Comments and Variance of Comment
Ratings: Video Data .......................................................................... 215 9.5 Remaining Issues in Quality Analysis of User-Generated Content ............. 215
9.5.1 Gold Standard and Standard Evaluation .......................................... 215 9.5.2 Privacy Issues on Metadata .............................................................. 216 9.5.3 Falsied/Spam Content ..................................................................... 216
9.6 Concluding Remarks .................................................................................... 217 References .............................................................................................................. 217
With the development of Web 2.0 technologies, public organizations regularly provide news, literature, or articles but at the same time common users also actively post and share their opinions or daily episodes through blogs, wikis, or other social media on the web. Thanks to the continuous growth of the volume of such user-generated content, users can easily access a wider variety of information nearly on any topic. For example, if we want to learn how to cook a particular dish, such as “Korean BBQ,” we can check out information about its numerous recipes on the web generated by other users. Such information is usually provided in diverse formats chosen by its authors. For instance, the authors may post recipes with photos they have taken. They may also post video so that the viewers may grasp the whole cooking process better because the video shows every step of it both in further detail and in a seamless way. User-generated content of this kind has the following benets. First, since the authors usually generate the contents from the reader’s (or viewer’s) perspective as much as possible, the contents can fully address the points that the readers may have most difculty in understanding. Second, the publishing time is comparatively quite short, so that the readers may get the most recent and relevant information. Third, the readers may receive feedback most readily from the authors via feedback sharing tools such as RSS or push notication services.
