neuralnetworks.calculate
delta
(delta input thetas activation-nodes output sigmoid-fn)
Calculate the ‘error’ (delta) from the expected input/output for the given thetas. This will also calculate the delta for each hidden layers.
forward-propagate
(forward-propagate input thetas sigmoid-fn)
Will return list of activation nodes for each theta
reshape-thetas
(reshape-thetas thetas thetas-dimensions)
Reshapes theta vectors based on the given thetas-dimensions where thetas-dimensions are vector of matrices dimensions