International Journal of Computer Science Issues

More than a traditional journal...

This paper aims to reveal the significant features of Amazon Mechanical Turk (MTurk) from a Mass Collaborative Knowledge Management perspective and to provide some recommendations for the improvement of that platform and the similar practices. This research mainly employs three types of research methodologies: Literature study was conducted to describe the RD of MTurk and its applications; Case study was conducted for examining its knowledge processing features form a mass collaboration perspective; Software engineering methodology is applied to discuss the underlying principles and application interfaces of that platform. MTurk provided better practices for Mass Collaborative Knowledge Management in terms of harnessing long tail side human intelligence, discomposing knowledge tasks, keeping track of users working history, providing qualification approaches, building a trusted third-party collaboration platform and introducing a pay-after-go payment approach. However, it has its limitations such as lacking a direct collaboration between the participants on a given task, machine unreadable knowledge presentation and failing to control the quality of knowledge processing. We can remedy those limitations by facilitating collaboration between the long tail-side workers, encouraging workers to submit extra data such as proof data, background information and relevant knowledge, introducing the Semantic Web into it and implementing man machine collaboration.