移动群智感知中基于用户偏好的参与者选择

Participant selection based on user preference in mobile crowdsensing

为了选择合适的参与者,提高用户提供的数据质量,结合用户偏好和任务特征,提出了新的参与者选择方法,通过用户偏好和任务特征的相似性作为用户对任务的感兴趣概率,同时平台根据历史信誉更新用户当前信誉.将用户信誉值设置为信誉得分,将用户对任务的感兴趣概率设置为兴趣得分,并将二者合并为用户选择得分.通过改进两阶段拍卖算法选择感兴趣概率高且信誉高的用户参与任务.在Python的实验环境中,通过改变参数设置,与已有的模型相比提高了数据质量,节省了预算.

In order to select appropriate participants,improve the quality of user-provided data,and combine user preferences and task characteristics,a new participant selection method was proposed. The similarity between user preferences and task characteristics was used as the probability of users’ interest in the task,while the platform updates userscurrent reputation according to historical reputation. The user’s credit value was set as the credit score,and the probability that the user was interested in the task set as the interest score,and the two were combined into the user’s selection score. By improving the twostage auction algorithm,users with high interest probability and high reputation are selected to participate in the task. In the experimental environment of Python,by changing the parameter settings,the quality was improved and the budget was saved compared to the existing model.