Post-Level Spam Detection for Social Bookmarking Web Sites

Author(s):  
Hsin-Chang Yang ◽  
Chung-Hong Lee
2014 ◽  
Vol 38 (6) ◽  
pp. 788-805 ◽  
Author(s):  
Duen-Ren Liu ◽  
Chuen-He Liou ◽  
Chi-Chieh Peng ◽  
Huai-Chun Chi

Purpose – Social bookmarking is a system which allows users to share, organise, search and manage bookmarks of web resources. However, with the rapid growth in the production of online documents, people are facing the problem of information overload. Social bookmarking web sites offer a solution to this by providing push counts, which are counts of users’ recommendations of articles, and thus indicate the popularity and interest thereof. In this way, users can use the push counts to find popular and interesting articles. A measure of popularity-based solely on push counts, however, cannot be considered a true reflection of popularity. The paper aims to discuss these issues. Design/methodology/approach – In this paper, the authors propose to derive the degree of popularity of an article by considering the reputation of the users who push the article. Moreover, the authors propose a novel personalised blog article recommendation approach which combines reputation-based group popularity with content-based filtering (CBF), for the recommendation of popular blog articles which satisfy users’ personal preferences. Findings – The experimental results show that the proposed approach outperforms conventional CBF, item-based and user-based collaborative filtering approaches. The proposed approach considering reputation-based group popularity scores on neighbouring articles indeed can improve the recommendation quality of traditional CBF method. Originality/value – The recommendation approach modifies CBF method by considering the target user's group preferences, to overcome the limitation of CBF which arises from the recommending only items similar to those the user has previously liked. Users with similar article preferences (profiles) may form a group of users with similar interests. A group's preferences may also reflect an individual's preferences. The reputation-based group preferences of the target user's group can be used to complement the target user's preferences.


Author(s):  
Jos van Iwaarden ◽  
Ton van der Wiele ◽  
Roger Williams ◽  
Steve Eldridge

The Internet has come of age as a global source of information about every topic imaginable. A company like Google has become a household name in Western countries and making use of its internet search engine is so popular that “Googling” has even become a verb in many Western languages. Whether it is for business or private purposes, people worldwide rely on Google to present them relevant information. Even the scientific community is increasingly employing Google’s search engine to find academic articles and other sources of information about the topics they are studying. Yet, the vast amount of information that is available on the internet is gradually changing in nature. Initially, information would be uploaded by the administrators of the web site and would then be visible to all visitors of the site. This approach meant that web sites tended to be limited in the amount of content they provided, and that such content was strictly controlled by the administrators. Over time, web sites have granted their users the authority to add information to web pages, and sometimes even to alter existing information. Current examples of such web sites are eBay (auction), Wikipedia (encyclopedia), YouTube (video sharing), LinkedIn (social networking), Blogger (weblogs) and Delicious (social bookmarking).


2011 ◽  
Vol 15 (3) ◽  
pp. 31-72 ◽  
Author(s):  
Toine Bogers ◽  
Antal van den Bosch
Keyword(s):  

In this paper, we depicts spam revelation, in perspective of the examination of posts, in social bookmarking districts. For consistent acknowledgment of spam posts, we propose a name estimation plot and a specific evaluation procedure for picking marks. The label estimation scores each tag. In the particular evaluation, the label scores in perspective of the utilization repeat and the degree of spammers are estimated and the thoughts of white tag and dim tag are introduced. Using these thoughts, names are proficiently arranged into the names demolishing the execution of spam revelation, the names pleasing in getting spammers, and the marks which should achieve a discipline. Finally, we propose semantic components to moreover upgrade the spam distinguishing proof


Sign in / Sign up

Export Citation Format

Share Document