In this project, I practice implementing neural network models from scratch using NumPy. I'm not going to expend too much effort creating the most functional classes, or optimizing the performance since I'm not trying to replace the current industry standard frameworks.
If you spot any critical errors, please reach out to me on LinkedIn.
Update: I added a neural network class for regression purposes from scratch.
The first neural network class I developed is a deep learning network with a single hidden layer. I downloaded the Bike Sharing Dataset from UCI and tested how the neural network performs. My main take-away from this exercise is not the bike rental predictions. My main take-away is that the model seems to be learning! Although the model is far from perfect, it performs significantly better than random guessing.
Upcoming: Sentiment Analysis Neural Network