| Method task | FPN | PNN [34] | SVM [11] |
| Network architecture | 7-3-3-1 | 1-40-4-3 | 1-3-3 |
| Training data | - | 40 Input-output pair 1: [Y, DDOS%] Input-output pair 2: [Y, DOS%] | |
| Activation function | Gaussian membership function (GMF): Equations (17), (22) | Gaussian Function [11] [34] | |
| Inference/ learning algorithm | Multivalued logic Max operation | ·Least mean square algorithm ·Gradient descent method ·Particle swarm optimization (PSO) algorithm | |
| Parameter assignment | ·Means and standard deviances of GMFs ·Weighted values, bfrom transitions to desired places ·Weighted values, Wfrom transitions to desired outputs | ·Population size: G = 20 - 40 [34] ·Inertia weight: w ·Acceleration parameters ·Uniformly random numbers ·Maximum number of allowable iteration: gmax, g = 1, 2, …, gmax = 50 | |
| Adjustable parameter | - | sk | tk, sk, biasconstrain: 0 < tk < 1 |
| Iteration training | - | < 50 | 20 - 40 |
| Convergent condition | - | MSEF < 10-2 | MSEF < 0.2 |
| Accuracy (%) | 95.0 | 95.0 | 65.0 |