Independent worker selection method based on community discovery and link prediction in crowdsourcing system

Abstract

The invention discloses an independent worker selection method based on community discovery and link prediction in a crowdsourcing system, and the method comprises the following steps: S1, worker historical data collection: collecting worker historical data on a crowdsourcing platform, carrying out the correlation analysis of the worker historical data, and finding the existing correlation betweenworkers; S2, analyzing independence of workers: analyzing existing relations among the workers by using the historical data of the workers and adopting a community discovery algorithm to obtain independence among the workers; S3, predicting the future independence of the workers, and predicting the future association possibility among the workers by adopting a link prediction technology; and S4,crowdsourcing worker selection. The method has the beneficial effects that the existing contact tightness and the future contact possibility of crowdsourcing workers are considered, the community discovery and link prediction technology is introduced to ensure the independence of the selected crowdsourcing workers, the possibility of plagiarizing and collusion among the workers is reduced, and thecompletion quality of crowdsourcing tasks is better improved.

Description

[0001] 【Technical field】 [0002] The invention relates to the field of crowdsourcing systems, in particular to a method for selecting independent workers based on community discovery and link prediction in the crowdsourcing system. [0003] 【Background technique】 [0004] When the crowdsourcing platform assigns tasks, it mainly considers factors such as workers' ability, reliability, and reputation. Through the combination optimization of workers and tasks and the introduction of incentive mechanisms, the benefits of tasks are maximized. However, existing works rarely consider the possible connections between users. Since quite a lot of crowdsourcing tasks need to be completed independently by workers, relying on the independent judgment of the public to ensure the effect of the result collection. Therefore, it is necessary to judge and screen the possibility of worker contact in task assignment, so as to ensure the individual independence of the selected worker set. At pres...

AI Technical Summary

Problems solved by technology

However, since the release and completion of crowdsourcing tasks are completed online through an open network, the publisher cannot understand the user's specific operations

However, the existing worker selection technology usually only focuses on factors such as the ability and historical records of a single or a combination of workers to complete the work, and does not consider the connection between workers, such as the similarity of the workers themselves, the conscious collusion of plagiarism between workers, etc. As a result, the collected workers have the same answers, plagiarism and other issues, which affect the final aggregation of crowdsourcing results

Images