Abstract:
Crowdsourcing is a process of judiciously selecting the right user (worker) from a large pool of online community who could solve the task. The undertaking of jobs by many online users (workers) simultaneously helps solving large-scale computational problems. After completion of the job, monetary reward would be offered to the user who has completed the job satisfactorily or efficiently. The main challenge in crowdsourcing platforms is to assign a task to a user as users are mainly available online and are unknown and unfamiliar to each other. To overcome this challenge, we have proposed a new algorithm for task assignment based on the trustworthiness of online users. Trustworthiness is calculated by using the belief and knowledge values (metrics) of individuals who have shown interest in taking up the task. We have used vector space model to find out that the person is knowledgeable to do the task at hand. The belief value for a user is calculated by using the reputation and the familiarity index of the user on the social media interactions. Tasks would be assigned to all the users whose trust value is above a specific threshold. We have evaluated the proposed algorithm on a social media application that we created for sharing expertise amongst the off-campus faculty colleagues of our university.