Simple Neural Network

Methods include:-
1. Computational Graph,
2. Shunting-yard Algorithm,
3. Forward Pass,
4. Dynamic Programming,
5. Memoization, and
6. Back Propagation.

Forward Pass:-
User gives liner or binomial equation with wights and biases. The Shunting-yard algorithm parsed the equation and created a computational graph. Then every node of the computational graph was differentiated using a hand-coded differentiation function. With the help of Dynamic Programming and user input (weights & biases), I calculated the value of every node inside the computational graph. Memoization helped the Dynamic Programming portion to be more efficient.
Subsequently, the output is calculated and compared with the actual output to compute the loss of the neural network using Euclidian Distance.