#Todo
- cross checking: running multiple train/test split runs to verify how stable the fit_error is
- multiclass classifier using NN
- normalizing input-score for linear regression
#Dependencies
- BGG XML API 2: https://boardgamegeek.com/wiki/page/BGG_XML_API2
- boardgamegeek API library (fork of libBGG): https://github.com/lcosmin/boardgamegeek, Documentation: http://boardgamegeek.readthedocs.io/
#Reads!
- B-Rank: https://arxiv.org/pdf/0908.2741.pdf (Marcel Blattner)
- http://bigdata-madesimple.com/k-nearest-neighbors-curse-dimensionality-python-scikit-learn/
- https://towardsdatascience.com/how-to-build-a-simple-song-recommender-296fcbc8c85
- Research: python n-dimensional space nearest neighbor
#Recommendation
- if I get more user-collections a collaborative filtering approach can be used - but this is not likely.
- Matrix Factorization: http://www.albertauyeung.com/post/python-matrix-factorization/
#Python Libraries
- boardgamegeek2
- tqdm
- sklearn
- numpy
- scipy