Given a datset of an ecommerce website collected from cengage.com, this project uses deep neural network with one hidden layer from scratch without the use of any neural network parameters library. The project aims in predicting the predictable user actions on the website.
The user actions based on the number given is as follows:
0 -> Bounce
1 -> Add to cart
2 -> Begin Checkout
3- -> Finish_Checkout
The algorithm's classification rate is 0.97 which is quite good.