“…For instance, Wang, et al [14] made evaluations on the ISCX VPN-Non VPN dataset in their study, but although the dataset used includes 206688 samples, there is also a difference in the number of classified applications. Similarly, Chiu, et al [17] made evaluations on the ISCX VPN-Non VPN dataset in their study, but the dataset used includes 222797 samples, but there is also a difference in the number of classified applications. Similar to these examples, when other studies such as Lotfollahi, et al [4], Bu, et al [8], Chen, et al [20], Izadi, et al [10] and Hu, et al [31] mentioned in Table 1 are examined, it will be seen that they have different dataset classes and sizes.…”