Name
Abstract | ||
|---|---|---|
The collaborative efforts of users in social media services such as Wikipedia have led to an explosion in user-generated content and how to automatically tag the quality of the content is an eminent concern now. Actually each article is usually undergoing a series of revision phases and the articles of different quality classes exhibit specific revision cycle patterns. We propose to Assess Quality based on Revision History (AQRH) for a specific domain as follows. First, we borrow Hidden Markov Model (HMM) to turn each article's revision history into a revision state sequence. Then, for each quality class its revision cycle patterns are extracted and are clustered into quality corpora. Finally, article's quality is thereby gauged by comparing the article's state sequence with the patterns of pre-classified documents in probabilistic sense. We conduct experiments on a set of Wikipedia articles and the results demonstrate that our method can accurately and objectively capture web article's quality. |
Year | DOI | Venue |
|---|---|---|
2011 | 10.1007/978-3-642-23091-2_50 | DEXA (2) |
Keywords | Field | DocType |
revision state sequence,revision history,quality class,revision phase,revision cycle pattern,different quality class,quality corpus,specific revision cycle pattern,web article,probabilistic quality assessment,wikipedia article | User-generated content,Data mining,Social media,Information retrieval,State sequence,Computer science,Probabilistic logic,Latent semantic analysis,Hidden Markov model,Database | Conference |
Citations | PageRank | References |
3 | 0.39 | 10 |
Authors | ||
3 | ||
Name | Order | Citations | PageRank |
|---|---|---|---|
| Jingyu Han | 1 | 16 | 4.67 |
| Chuandong Wang | 2 | 5 | 0.78 |
| Dawei Jiang | 3 | 380 | 21.67 |