This is the code for "Neural Networks - The Math of Intelligence #4" by Siraj Raval on Youtube

Create your own self organizing map implementation using numpy. Bonus points if you put your code in a Jupyter notebook and document your steps well. You can use any dataset you like, here is a good place to find some datasets. Post your Github link in the comments section of the video. Good luck!

This is the code for this video on Youtube by Siraj Raval as part of The Math of Intelligence Series. I go over 4 different neural networks in the video, and you can find them in this repository.

• numpy
• copy
• pil

Install dependencies using pip

The simple AF notebook contains the code for the network with the added hidden layer. For the simplest version, see this. The Recurrent notebook contains the recurrent code and the self organizing map contains the code for the unsupervised model.

Type jupyter notebook in the terminal in the main directory and the code will appear in your browser. Install jupyter from here if you haven't yet.

Credits go to Trask and Litery. I've merely created a wrapper to get people started.

Install Conda https://conda.io/docs/installation.html

//OSX / Linux Windows conda create -n maths python=3.5 source activate maths conda install pandas matplotlib jupyter notebook scipy scikit-learn nb_conda