A Novel Single Neuron Perceptron with Universal Approximation and XOR Computation Properties

Algorithm 1

Proposed SNP algorithm.

Input: Initial random weights; 𝐰_𝟏,𝐰_𝟐,…,𝐰_𝐧,𝐰_𝐧+𝟏 and input bias b

(1) Take 𝑘th learning sample (𝑘th 𝑝 and 𝑇)

(2)𝑝_𝑛+1=max_{𝑗=1,…,𝑛}(𝑝_𝑗)

(3) Calculate the final output 𝐸_𝑜 and error

     𝐸_𝑜=tansig∑^𝑛+1_𝑗=1𝑤_𝑗×𝑝_𝑗+𝑏

     𝑒=𝑇−𝐸_𝑜

(4) Update the weights by using excitatory rule

     𝑤_𝑗=𝑤_𝑗+𝛼max(𝑒,0)𝑠_𝑗; for  𝑗=1,…,𝑛+1

     𝑏=𝑏+𝛼max(𝑒,0)

(5) Update the weights by using inhibitory rule

     𝑤_𝑗=𝑤_𝑗−𝛼max(−𝑒,0)𝑠_𝑗; for  𝑗=1,…,𝑛+1

     𝑏=𝑏−𝛼max(−𝑒,0)

(6) If 𝑘<number of training patterns then 𝑘=𝑘+1 and proceed to the first

(7) Let epoch = epoch + 1 and 𝑘=1

(8) If the stop criterion has not satisfied proceed to the first