Fig. 10
From: Improving MapReduce privacy by implementing multi-dimensional sensitivity-based anonymization
Comparison between MDSBA and MDTDS with \(\overline{k}\) and data size variations. This diagram illustrates our MDSBA performance, in comparison with the To-Down specialization method. MDSBA processing time is a prominent in small values of \(\overline{k}\). However, MDTDS has lower latency on larger values of k, while MDSBA consumed the same time in both values of \(\overline{k}\), when \(\overline{k} = 50\), or \(\overline{k} = 250\). This fixed timestamp refers to the time consumed on applying the Obvious Guess. Larger values of \(\overline{k}\) requires less time spent on Obvious Guess processing, and more time on anonymization processing, and vice versa. Hence, the total of both processing times remains close similar for all \(\overline{k}\) values