Determining the currency of dynamic data

Xiaoou Ding, Hongzhi Wang, Yitong Gao, Jianzhong Li, Hong Gao

Abstract: Data currency is one of the most important factors influencing the data quality, which aims at solving the entities' attributes currency decision, error detection and repairing problems, with timestamps incomplete or not exist. However, with the rapid growth of volume and updating of requirements on data, serial processing methods cannot reach the best effect of data currency determining. That motives us to propose methods of currency determining on dynamic data in this paper. The structure of entity query B-Tree and entity storage static linked list are established to improve querying and updating both data currency graph and currency scores. With the analysis on the influence of four main parameters of the algorithms, experimental results on both real-life and synthetic data verify the efficiency of our methods.