| Crowdsensing network has the characteristics of low data collection cost,rich data types and large collection scale,which has been widely concerned by researchers and service providers.The participant is the provider of sensing data,which is an important key to the high quality and efficiency of crowdsensing.Therefore,It has important research significance for driving more and better participants to participate in sensing tasks,collect data,and contribute their own strength.Effective participant privacy preservation can reduce privacy violations and increase the enthusiasm for task participation.The increase in the number of participants also has a positive impact on the sensing data quality.In view of the current research that participants’ personalized privacy preservation needs cannot meet and the balance of privacy and data utility are difficult to find,participants’ personalized privacy preservation scheme was proposed with improved Local Differential Prvacy based on privacy and utility tradeoff.The proposed scheme will maximize data availability while fully meeting participants’ privacy needs.It makes that participants have more trust in crowdsensing,eliminates the worries of participants,and encourages more participants to participate sensing tasks.However,the expansion of the number of participants will lead to the decline of the quality of participants.Reasonable selection of participants can improve sensing quality.Participants’ optional selection scheme is proposed based on hybrid recommendation algorithm based on collaborative filtering and association rules.Based on the performance of the participants in different tasks,the collaborative filtering algorithm was applied to select participants to improve the sensing quality of the sensing tasks.On this basis,task attributes are introduced to improve the optimization scheme,which solves the problem of poor selection effect of participants under the new task.Association rule mining is used to optimize similarity calculation to further solve the impact of task-attribute matrix data sparsity,and improve the accuracy of participant selection.The participants’optional selection scheme guides crowdsensing to select better participants accurately,drive better participants to participate in the sensing task,and improve the sensing quality. |
