I am a 20 something physicists trying to learn machine learning. When I am faced with a difficult problem I find it usefull to write a text about it. This is where I keep these texts. They are mostly here so future Mikkel can find them easily but feel free to read along.
Lists and Vectors: I have been running into trouble while implementing machine learning algorithms in python -- not because the math is giving me a hard time, but because I unintentionally mix up the different structures that Python and the Numpy package provide. In this post I attempt to clear the waters.
One of my favortie features of Spotify, Steam and Goodreads is the reccomendations they provide. So I've decided the write a reccomender system for video games.
Collaberative Filtering - The mathematics: Before writing any code one should first understand the algorithm and the mathematics behind it. This post is all about the mathematics.
Collaberative Filtering - Writing it in python: While the level of abstraction mathamatics providde is great any implementation often raises several issues. In this post I go through the particular implementation I have written in Python.
Collaberative Filtering - Testing the model: Once you have a working system it needs to be tested. This take a lot of work and in many ways this is where it gets interesting -- how can you tweak the parameter to make it preform better? Is is baiased in any way? These are the kinds of questions that I address in this post.