Decision Support System for Alarm Correlation in GSM Networks Based on Artificial Neural Networks

Table 3

Training experiments.

Number	Type of Network	Inputs	Outputs	Training Alg.	Hidden layer	Hidden neurons
1	Feed-Forward	10	10	Trainlm	1	10
2	Feed-Forward	10	10	Trainscg	1	10
3	Feed-Forward	10	10	Trainrp	1	10
4	Feed-Forward	10	10	Trainlm	1	20
5	Feed-Forward	10	10	Trainscg	1	20
6	Feed-Forward	10	10	Trainrp	1	20
7	Feed-Forward	10	10	Trainlm	2	10
8	Feed-Forward	10	10	Trainscg	2	10
9	Feed-Forward	10	10	Trainrp	2	10
10	Feed-Forward	10	10	Trainlm	2	20
11	Feed-Forward	10	10	Trainscg	2	20
12	Feed-Forward	10	10	Trainrp	2	20
13	Cascade-Forward	10	10	Trainlm	1	10
14	Cascade-Forward	10	10	Trainscg	1	10
15	Cascade-Forward	10	10	Trainrp	1	10
16	Cascade-Forward	10	10	Trainlm	1	20
17	Cascade-Forward	10	10	Trainscg	1	20
18	Cascade-Forward	10	10	Trainrp	1	20
19	Cascade-Forward	10	10	Trainlm	2	10
20	Cascade-Forward	10	10	Trainscg	2	10
21	Cascade-Forward	10	10	Trainrp	2	10
22	Cascade-Forward	10	10	Trainlm	2	20
23	Cascade-Forward	10	10	Trainscg	2	20
24	Cascade-Forward	10	10	Trainrp	2	20