众包质量控制策略及评估算法研究

摘要

With the rapid development of Internet technology,crowdsourcing service as a flexible and effective solution,more and more people begin to be concerned about the matter.Crowdsourcing applications generate complex background for online network trading platform,and quality control is the key of crowdsourcing applications.So how to effectively improve the quality of task and identify malicious workers is becoming a hot topic in research community.We try to judge the worker whether good enough by evaluating their answer for the task.This paper proposes a new crowdsourcing quality assessment model.The real experiment results shown that the method could achieve good performance.%随着Internet技术的快速发展,众包作为一种灵活有效的解决问题方式,开始受到人们越来越多的关注.由于众包的自由松散组织模式,使得如何有效地控制任务完成质量,并将欺骗类型工作者识别出来,成为目前众包研究中一个急需解决的问题.文中基于对众包工作者提交结果的评估与分析,提出了一种阶段式的动态质量控制策略,同时给出了一个组合式众包结果质量评估方法框架.经过实际数据的测试,文中提出的质量控制策略和众包结果质量评估方法具有较好的效果.