Research On Crowdsourcing Task Assignment Mechanism For RDF Data Quality Evaluation

With the explosive growth of RDF data on the network,the quality of data is becoming more serious.Inaccuracies,inconsistencies,or incompleteness of the underlying data severely affect the use of RDF data.The current main means of solving these difficultto-detect or repair RDF data quality problems is still the expert intervention,which is extremely costly.In order to meet the challenge of RDF data quality assessment,crowdsourcing has become a powerful mechanism to effectively solve the problem of machine problems with the wisdom of the public.However,crowdsourcing workers have different knowledge backgrounds,the quality of their submitted crowdsourcing results is bound to be uneven.In order to effectively improve the quality of RDF data,it is necessary to control the quality of crowdsourcing workers and their submission of crowdsourcing task results.This paper has conducted in-depth research based on the above problems.The main work is divided into two parts:(1)A concept professional assessment model is proposed to select high-quality professional crowdsourcing workers to complete the RDF data crowdsourcing tasks.The model extracts test task instances from the standard knowledge base according to the concept hierarchy tree of the RDF data crowdsourcing task.And use the knowledge representation method to automatically build a set of options for the test task instance to generate a concept professional test task.(2)Propose a dynamic assignment strategy for crowdsourcing tasks based on concept professionalism,which is used to match the most suitable RDF data crowdsourcing tasks for workers.The strategy builds a worker's conceptual accuracy vector based on a conceptual professionality model.Combined with the worker's conceptual distance and conceptual accuracy,it dynamically matches the nearest RDF data crowdsourcing task,so that each worker is assigned to a crowdsourcing task that matches his or her own knowledge.The experimental results show that the concept professional assessmemnt model proposed in this paper can effectively select high-quality professional workers according to the concept hierarchy tree of RDF data crowdsourcing tasks;and the task assignment strategy based on concept professionalism effectively improves the quality of crowdsourcing tasks.