Social-Network-Assisted Task Recommendation Algorithm in Mobile Crowd Sensing

With the popularity of mobile smart devices, collecting data has become more convenient. Freeing from the constraints of professional equipment, Mobile Crowd Sensing (MCS) has gained wide attention. As the key component of MCS system, task recommendation directly affects the quantity and quality of task completion. However, most of previous task recommendation modules in MCS system only consider the situation where users complete tasks independently, without further consideration of the possibility that users can seek assistance through social networks. In this paper, we come up with a task recommendation algorithm combined social networks to maximize the number of completed tasks. We build up a user-task rating matrix based on the number of tasks performed by each user, and then we use the matrix factorization method to get the latent factor matrix. According to the latent factor matrix, we greedily select some tasks for each user. Next, we calculate the user extroversion and intimacy with others through social networks data to get the probability of users asking for help from their friends. We get the scoring matrix and task recommendation list, considering that users could complete the task together. Finally, we conduct lots of experiments based on a real-world dataset, and the experimental results show that our solution outperforms the existing algorithms.