Implementations of some Machine learning models.

End to end deep neural network implementation from scratch.

1. This implementation uses only sigmoid activation
2. Optimization techniques implemented are classic gradient descent, gradient descent with momentum, RMSProp and Adam optimization. All of these are mini-batched by default
3. The deep neural network is implemented only with L2 - Regularization

Note: The notebook has many similarities to the programming assignments from Coursera Deep learning courses because that is where I learned it.

Collaborative filtering from scratch.

Implementation of recommender system using Collaborative filtering algorithm by optimizing a single cost function to learn both User features and Product features.