Re-ranking in user-driven reputation systems with splay tree. (2018)

• Record Type: Journal Article
• Authors: Jayashree, R.; Christy, A.
• Is Part Of: International journal of material & product technology. Volume 56:Number 1/2(2018)
• Page Start: 3
• Page End: 22
• Publication Date: 2018
• Subjects: adaptive and intelligent educational systems -- user-driven reputation system -- reputation management -- ranking using a splay tree -- ranking algorithm -- question-answering websites -- re-ranking -- semi-splay -- Zipf's or power law; Manufacturing processes -- Periodicals; Materials -- Testing -- Periodicals; 620.11
• Languages: English
• ISSNs: 0268-1900

Abstract

In e-learning environments, trust relationship plays a vital role in establishing collaborative activities among co-learners. Reputation is a fundamental source for measuring co-learners' trust. An efficient and effective way to construct trust relationship among faculties and students in e-learning environment is ranking. User-driven reputation systems are based only on the feedback or ratings provided by the users. Users with higher points obtain high rating compared to less scored users. Thus, by Zipf's law, alleged low users are pushed to the bottom of the ranking list. This condition is avoided by encouraging less likely users and preventing them from moving further down in ranking level. Low ranked users are provided with few more chances to participate actively in the e-learning environment. A splay tree is a binary search tree with self-balancing skill. The splay tree brings the recently accessed item to the top of the tree. A splay tree is used to represent user's ranks, and low ranked users are semi-splayed again in the tree thus preventing them from further drowning in the ranking list for the fixed number of times. In this paper, we presented Rank_Improvement algorithm to enhance average scored users and compared with the ranking algorithm in the existing question-answering websites.