Efficiently Preserving Privacy on Large Trajectory Datasets
Chen-Yi Lin, Yuan-Chen Wang, Wan-Tian Fu, Yunsheng Chen, Kuan-Chen Chien, Bing-Yi Lin · 2018
Modern people use credit cards, RFID cards, and so on for transaction payment, and these transaction trajectories will be input to a dataset. If the dataset with user privacy is released, attackers can use partial trajectory knowledge as a quasi-identifier. To protect user privacy, we propose a trajectory anonymous method to avoid attackers attacking the original dataset randomly, and suppress part of the trajectories, so that the attackers cannot correctly link the correct trajectory with sub-trajectory knowledge. In addition, we use a tree-based architecture, which greatly improves computational efficiency and reduces computational time. Furthermore, we simplify the formula for computing information loss to improve the efficiency of the computation. Finally, we demonstrate the effectiveness of the proposed algorithm in data utility preservation and efficiency through experiments.