Towards Tax Evasion Detection Using Improved Particle Swarm Optimization Algorithm

Table 2

Initialization of parameters.

Parameters	Value	Algorithm
Number of particles	50	IPSO
Iteration	500	IPSO
C1	1	IPSO
C2	2	IPSO
	1	IPSO
Neuron input layer	8	MLP
Neuron hidden layer	10	MLP
Number of hidden layers	3	MLP
Training data	80%	MLP
Testing data	20%	MLP
Activation function	Sigmoid	MLP
Learning rate	0.15	MLP
C	1–10	SVM
	0–1.0	SVM