Impact of Parameter Tuning for Optimizing Deep Neural Network Models for Predicting Software Faults

Table 11

Experiment performed on the PC2 dataset by tuning the parameter i.e., learning rate = -5, epochs = 100, 200, 300, 500, 1000, and 2000, number of layers = 5, and dropout = 0.2 and 0.5.

No. of layers	5	5	5	5	5	5
No. of neurons with dropout	80 (0.2), 80, 80, 80, and 200	80 (0.2), 80, 80, 80, and 200	80 (0.2), 80 (0.2), 80, 80, and 200	80 (0.5), 80 (0.5), 80 (0.5), 80 (0.5), and 200 (0.5)	80 (0.5), 80 (0.5), 80 (0.5), 80 (0.5), and 200 (0.5)	80 (0.5), 80 (0.5), 80 (0.5), 80 (0.5), and 200 (0.5)
Lr	1.00E − 05	1.00E − 05	1.00E – 05	1.00E − 05	1.00E − 05	1.00E − 05
Epoch	100	200	300	500	1000	2000
Accuracy	0.9623	0.9623	0.9623	0.9623	0.9623	0.9623
Loss	0.5104	0.4097	0.3033	0.1055	0.1112	0.2213