Independent Worker Selection Algorithm:Based On Community Detection And Link Prediction

The crowdsourcing model benefits from the rapid development of the Internet,which maximizes the efficiency and quality of idle resources and unmet needs in the real world through the Internet.A variety of tasks are not limited to the offline environment or isolated task groups,but break through the limitations of time and space and organizational structure,forming a collaborative form of work between different individuals.However,there are also some problems in the crowdsourcing mode,such as plagiarism of workers' tasks,which greatly affects the quality of the completion of crowdsourcing tasks.This paper proposes the choice of crowdsourcing workers based on independence,which effectively avoids my situation?This paper proposes an innovative independent worker selection algorithm by referring to the technology of community discovery and link prediction.The specific research contents,contributions and innovations are as follows:1)The homogeneity analysis of crowdsourcing workers.In the existing crowdsourcing network worker selection algorithm,the corresponding worker selection is mainly based on the credibility,accuracy,and bidding of the worker.However,due to the continuous development of the crowdsourcing model,the connection between workers has been continuously enhanced,the homogeneity problem has become increasingly prominent,and the plagiarism of tasks has become increasingly serious.This paper analyzes the use of some homogeneity indicators for the workers' group.proves the homogeneity between workers,and discovers the impact of some characteristics on workers' homogeneity such as location,time,and social relationships.It provides a theoretical basis for workers to choose their independence.2)Independent worker selection based on modularity optimization.To achieve independent worker selection,we apply the theory of community detection.Community detection technology can better classify more connected workers into the same community,so the associations of workers in different communities are correspondingly less.This paper proposes a community discovery expansion algorithm based on modularity optimization,and selects workers in different communities through multiple selection methods to select a set of candidate workers with higher independence.3)Independent worker selection based on link prediction.Community discovery technology can only perform corresponding analysis based on existing relevant data.However,there are a large number of unknown features and associations in reality.In order to reduce the cost of the experiment and predict the possible correlation of workers in the future,this paper further proposes an independent worker selection algorithm based on similarity link prediction.The similarity between workers is calculated by several similarity indicators,and the corresponding similarity matrix is established,so that some workers with the lowest overall similarity are selected from the set of candidate workers.Finally,the appropriate similarity index is determined by the analysis of worker homogeneity.