An Overview of Query Processing on Crowdsourced Databases

Abstract

Crowd-sourcing is a powerful solution to correctly answer expensive and unanswered queries in the database. This includes queries on a database with uncertain and incomplete data. The crowd-sourcing attempts to exploit human abilities to process these difficult tasks and workers helped to provide accurate results utilizing the available data in the crowd. The crowd-sourcing database systems (CSDB) combined the ability of the crowd with the relational database by using some variant of the relational database with minor changes. This paper surveys and examines the leading studies conducted on the area of query processing for both traditional and preference queries in crowd-sourcing databases. The focus is given on highlights the strengths and the weakness of each approach. A detailed discussion for the current and future trends research relevant to query processing in the area of crow-sourced databases is also demonstrated.

CIDS: An Efficient Algorithm for Processing Skyline Queries for Partially Complete Data in Cloud Environment

“…In the last part of the experiment set, the amount of data transferred from one data center to another is measured. It is proven from the literature that these parameters are the most crucial parameters that influence skyline query processing [1], [2], [4], [13], [23], [36], [44], [45].…”

“…Among preferences queries, the skyline query is considered one of the best queries when it comes to user preference. These queries have been intensively used in many databases such as multi-criteria decision making, decision support [7][8], crowdsourcing databases [9]- [13], hotel recommender [14], restaurant finder [2][14], temporal databases [15], and cloud databases [16]. Since the skyline operator was introduced, the researchers have been working on improving it, in terms of efficiency and accuracy.…”

IDSA: An Efficient Algorithm for Skyline Queries Computation on Dynamic and Incomplete Data With Changing States

“…These definitions and notations assist in clarifying the proposed approach. (7,5,1,8), pi ≻ pj as pi is better than pj in the second and third dimensions (true on the second part of the condition), while pi is equal to pj in the first and fourth dimensions (true on the first part of the condition). This definition applies to those tuples that are comparable (refer to Definition 2).…”

“…Throughout this paper, we assume that bigger values are preferred over smaller ones. For instance, consider the tuples pi(7,6,2,8) and pj…”