摘要: In this chapter, we discuss detailed statistics of the popular Amazon Mechanical Turk (AMT) marketplace to provide insights in task properties and requester behavior. We present a model automatically infer communities based on keywords. Hierarchical clustering is used identify relations between keywords associated with tasks. novel techniques rank requesters by using graph-based algorithm. Furthermore, introduce models methods for discovery relevant crowdsourcing brokers who are able act as intermediaries platforms such AMT.