Planar data classification with one hidden layer

This is a programming assignment of an online course to show how well a shallow neural network can perform on planar data classification over logistic regression. In this assignment, I implemented forward and backward propagations with tanh and sigmoid activation funtions.