Please enable JavaScript.

Coggle requires JavaScript to display documents.

Artificial Neural Network (Back-Propagation Network (Steps (1. Forward …

- - - - Net Input of a Node
        \( Net_{h1/o1} = \sum_{i=1}^{N} w_ix_i + x_0 \)
      - Output of a Node
        \(Out_{h1/o1} = \frac{1}{1 + e^{-Net}} \)
      - Calculate Total Error
        \( E_{total} = \sum_{i=1}^{N} \frac{1}{2} (target-output)^2 \)
    - - \(\Delta w_k = ABC \)
        \( A = I_k \)
        \( B = f(net)(1 - f(net)) \)
        \( C = target-output \)
      - Output Layer
        \( \Delta w_{j,k} = I_k [f(net_{j,k})(1-f(net_{j,k}))] (T_k - O_k) \)
        
        j = source hidden node
        
        k = output node
      - Hidden Layer
        \( \Delta w_{i,j} = I_k [f(net_{i,j})(1-f(net_{i,j}))] C_j \)
        \( C = \sum_k w_{j,k} [f(net_{j,k})(1-f(net_{j,k})] (T_k - O_k) \)
        
        i = source hidden node
        
        j = dest hidden node
        
        k = output node