neural network

I created a feed-forward backpropagation network neural network using an arduino. In the backpropagation algorithm, the network starts with random weights. These weights are adjusted based on the training sets of output and inputs. This network has three layers: the input layer, the hidden layer, and the ouput layer. The input layer is the data that is entered, and the output is the data that is being solved. The hidden layer allows for the data to be solved, which in this case is a simple XOR truth table.

Code:

variables:

• PatternCount: # of training items or rows in the truth table.
• InputNodes: # input neurons.
• OutputNodes: # output neurons.
• HiddenNodes: hidden neurons.
• LearningRate: changes how much of the error is actually backpropagated.
• Momentum: changes how the results of the previous iteration affect the current iteration.
• InitialWeightMax: initializes maximum starting value for weights.
• Success: The threshold for error at which the network will be said to have solved the training set.

Output: