Social tags take an important role in exploratory search. In collaborative tagging systems, users are allowed to annotate resources with tags. The significant challenges in such systems are the uncertainty of tag quality and the incomplete annotation on a large number of resources. Based on the observation that these problems can be statistically negligible after receiving sufficient tags, we propose a novel incentive mechanism to reward taggers according to the quality of their bookmarks, called the Quality-based dynamic Incentive Mechanism (QIM). To well evaluate the quality of bookmarks, we design some quantitative evaluation methods. The reward allocation function is proposed to allocate the budget to different taggers based on their bookmark quality and the tagging states of annotated resources. We perform experiments to evaluate our method on three public datasets collected from real tagging systems. Comparing with previous works, the adopted principle of ``high quality deserves high price'' in this paper can encourage users to annotate seriously. The experimental results show that our method gets higher tagging quality of resources under a fixed budget. Moreover, it requires less time and less money to achieve the stable tagging state of a system.
%0 Journal Article
%1 Xu2015
%A Xu, Haoran
%A Zhou, Dandan
%A Sun, Yuqing
%A Sun, Haiqi
%D 2015
%J Distributed and Parallel Databases
%K Quality-based_dynamic_Incentive_Mechanism crowd_sourcing folksonomy information_access tag_quality
%N 1
%P 69--93
%R 10.1007/s10619-014-7164-8
%T Quality based dynamic incentive tagging
%U http://dx.doi.org/10.1007/s10619-014-7164-8
%V 33
%X Social tags take an important role in exploratory search. In collaborative tagging systems, users are allowed to annotate resources with tags. The significant challenges in such systems are the uncertainty of tag quality and the incomplete annotation on a large number of resources. Based on the observation that these problems can be statistically negligible after receiving sufficient tags, we propose a novel incentive mechanism to reward taggers according to the quality of their bookmarks, called the Quality-based dynamic Incentive Mechanism (QIM). To well evaluate the quality of bookmarks, we design some quantitative evaluation methods. The reward allocation function is proposed to allocate the budget to different taggers based on their bookmark quality and the tagging states of annotated resources. We perform experiments to evaluate our method on three public datasets collected from real tagging systems. Comparing with previous works, the adopted principle of ``high quality deserves high price'' in this paper can encourage users to annotate seriously. The experimental results show that our method gets higher tagging quality of resources under a fixed budget. Moreover, it requires less time and less money to achieve the stable tagging state of a system.
@article{Xu2015,
abstract = {Social tags take an important role in exploratory search. In collaborative tagging systems, users are allowed to annotate resources with tags. The significant challenges in such systems are the uncertainty of tag quality and the incomplete annotation on a large number of resources. Based on the observation that these problems can be statistically negligible after receiving sufficient tags, we propose a novel incentive mechanism to reward taggers according to the quality of their bookmarks, called the Quality-based dynamic Incentive Mechanism (QIM). To well evaluate the quality of bookmarks, we design some quantitative evaluation methods. The reward allocation function is proposed to allocate the budget to different taggers based on their bookmark quality and the tagging states of annotated resources. We perform experiments to evaluate our method on three public datasets collected from real tagging systems. Comparing with previous works, the adopted principle of ``high quality deserves high price'' in this paper can encourage users to annotate seriously. The experimental results show that our method gets higher tagging quality of resources under a fixed budget. Moreover, it requires less time and less money to achieve the stable tagging state of a system.},
added-at = {2017-04-12T20:51:06.000+0200},
author = {Xu, Haoran and Zhou, Dandan and Sun, Yuqing and Sun, Haiqi},
biburl = {https://www.bibsonomy.org/bibtex/2613c890400d2d25981b6323554eca742/charcharbinx},
doi = {10.1007/s10619-014-7164-8},
interhash = {f6f7a29325a39f52ce0fc0307a7b06c4},
intrahash = {613c890400d2d25981b6323554eca742},
issn = {1573-7578},
journal = {Distributed and Parallel Databases},
keywords = {Quality-based_dynamic_Incentive_Mechanism crowd_sourcing folksonomy information_access tag_quality},
number = 1,
pages = {69--93},
timestamp = {2017-04-13T01:57:54.000+0200},
title = {Quality based dynamic incentive tagging},
url = {http://dx.doi.org/10.1007/s10619-014-7164-8},
volume = 33,
year = 2015
}