MC-MLDCNN: Multichannel Multilayer Dilated Convolutional Neural Networks for Web Attack Detection

Table 2

Vocabulary table of characters and their indices in the one-hot encoding matrix.

Column	Index
abcdefghijklmnopqrstuvwxyz	1–26
0123456789	27–36
-,;.!?:’ ”/∖ | _@#$%ˆ&∗∼‘+-=<>()[]{}	37–69
UNK	70
PAD	0

It contains frequently used alphanumeric and special characters in the requests (indices 1–69). Any other character is defined as an unknown character (UNK with index 70). The 0 index is used for padding purpose.