Abstract
This paper proposes a way to filter out the people who give distorted movie ratings in On-line environment. With the movie rating data from Naver (www.naver.com), we have found two typical patterns that the malicious movie raters have presented. The first is that their points are very extreme. The second is that their opinions with regard to the movie are normally very irrelevant to the content of the movie. Leveraging this fact, the proposed scheme could filter out around 7～20% of the raters, which leads to non-trivial change of the ranking of the movies. This paper contributes to the people who want to know the true ratings of movies.
| Translated title of the contribution | Detection of Malicious Rate Evaluation and Prediction of True Rate in Movie Rating |
|---|---|
| Original language | Korean |
| Pages (from-to) | 213-218 |
| Number of pages | 6 |
| Journal | 정보과학회 컴퓨팅의 실제 논문지 |
| Volume | 20 |
| Issue number | 4 |
| State | Published - 2014 |