【摘要】 在众包质量评估中使用EM算法时较多赋予工作者相同的权重将导致结果容易陷入局部最优,然而工作者提交答案的可信度不尽相同,信誉良好的工作者给出的答案可信度更高;为此提出一种考虑工作者信誉的众包质量评估方法。首先根据工作者在众包平台上的历史交易行为建立工作者信誉模型,然后将信誉作为评价标准对工作者进行筛选,并以工作者信誉值为权重加入到EM算法,以改进其初始值确定过程,改善评估结果。进一步通过实例验证了该方法的有效性,并对提出方法的阈值参数进行了率定。更多还原

【Abstract】 When the EM algorithm is used in crowdsourcing quality assessment, the same weight will result in the result falling into local optimum easily. However, the reliability of the answer submitted by the worker is not the same, and the credibility of the answer given by the worker with good reputation is even higher. In this paper, a method of crowdsourcing quality evaluation considering the reputation of workers is proposed. Firstly, according to the worker’s historical transaction behavior on crowdsourcing platform, the reputation model of the worker is established; then the reputation is selected as the evaluation criterion, and the value of the worker’s reputation is added to the EM algorithm, in order to improve its initial value determination process, and improve the evaluation results. An example is given to verify the effectiveness of the proposed method, and the threshold parameters of the proposed method are determined.更多还原