PROJECT TITLE :

Reducing Uncertainty of Probabilistic Top-k Ranking via Pairwise Crowd sourcing - 2017

ABSTRACT:

Probabilistic top-k ranking is an important and well-studied question operator in uncertain databases. However, the standard of high-k results might be heavily littered with the ambiguity and uncertainty of the underlying data. Uncertainty reduction techniques have been proposed to improve the quality of high-k results by cleaning the initial information. Unfortunately, most knowledge cleaning models aim to probe the precise values of the objects individually and so don't work well for subjective knowledge varieties, like user ratings, which are inherently probabilistic. In this paper, we have a tendency to propose a novel pairwise crowdsourcing model to scale back the uncertainty of high-k ranking employing a crowd of domain specialists. Given a crowdsourcing task of limited budget, we have a tendency to propose economical algorithms to pick the best object pairs for crowdsourcing that will usher in the best quality improvement. Intensive experiments show that our proposed solutions outperform a random selection technique by up to thirty times in terms of quality improvement of probabilistic prime-k ranking queries. In terms of efficiency, our proposed solutions will reduce the elapsed time of a brute-force algorithm from several days to at least one minute.