Trust in Intrusion Detection Systems: An Investigation of Performance Analysis for Machine Learning and Deep Learning Models

Table 8

Feature selection of KDD dataset using values.

Feature No	Feature name	value	Statistically significant
1	Duration	2.33E − 60	Yes
2	Service	0.604708	Yes
3	src_bytes	1.34E − 68	Yes
4	Land	9.51E − 128	Yes
5	Urgent	0.402631	Yes
6	hot	5.01E − 230	Yes
7	num_compromised	2.69E − 43	Yes
8	su_attempted	2.45E − 21	Yes
9	num_file_creations	3.48E − 11	Yes
10	num_shells	9.85E − 43	Yes
11	num_access_files	6.50E − 26	Yes
12	num_outbound_cmds	0.043069	Yes
13	is_host_login	0.039867	Yes
14	srv_diff_host_rate	1.56E − 70	Yes
15	protocol_type	0	No
16	Flag	0	No
17	dst_bytes	0	No
18	wrong_fragment	0	No
19	num_failed_logins	0	No
20	logged_in	0	No
21	root_shell	0	No
22	is_guest_login	0	No
23	Count	0	No
24	serror_rate	0	No
25	rerror_rate	0	No
26	same_srv_rate	0	No
27	diff_srv_rate	0	No
28	dst_host_count	0	No
29	dst_host_diff_srv_rate	0	No
30	dst_host_srv_diff_host_rate	0	No