Novel Approaches for Enhancing Crowd Sourcing Service

Abstract

ABSTRACT newline newline newline newline newlineCrowdsourcing can be utilized to integrate the computation performed by human into a several of data exhaustive procedures that are complicated for computers alone to resolve it. By using crowd powered algorithms, the human can be considered as processing elements. But in case of crowdsourcing, human is considered as a data source. The major objective of this research is to develop a novel system for crowdsourcing service. Query optimizer and crowdsourcing executor are the two main process in crowdsourcing services. Hence there is a need for developing suitable and better strategies for these two processes. In order to provide efficient crowdsourcing system, three approaches are used, the first two approaches are aimed to develop a better query optimizer. The soft computing techniques are adopted to achieve better performance in query optimization. Then the third approach is targeted to develop a better crowdsourcing executor. It is essential to enhance the performance in crowdsourcing task management. Hence, from these three approaches, we can obtain a novel and better system for crowdsourcing service. newlineFor query optimization, the Structured Query Language (SQL) query obtained from the user and the cost budget is given as an input and an optimized query plan is generated with cost based crowd optimization. Two performance concerns in crowdsourcing system are latency and cost. Query optimizer obtain the user query and also parses an SQL query into a tree structure query plan and applies optimization strategies to the initial plan. There are three kinds of crowd powered operators namely cjoin, cselect and cfill. These operators are applied for query evaluation and the operators are executed in iterative procedure with bottom up traversal of tree planning. newlineIn order to generate Human Intelligence Tasks (HITs) for publishing it in crowdsourcing platform, crowdsourcing task manager is employed with crowdsourcing executor. For optimization cost minimization and cost bounded latency minimization are t