This repository holds all relevant code to evalaute recsys algorithms over time. The main code is in the file lenskit_gs_main.py. The algorithms and evaluation are from the Lenskit libary.
Dependencies:
- pandas
- numpy
- sklearn
- tensorflow
- lenskit
The following datasets were used:
- Neflix
- Amazon (books, instant video, toys and games, Music)
- Movielens (100k, 1M, 10M)
- yelp Any other datasets that include a timestamp can be used. The reading instructions just need to be added to the function 'read_datasets' in the util.py file.