Instructions to run this code:
-
Place in the "./data/input" folder the following files:1
- item_data.jl
- test_dataset.jl
- train_dataset.jl
-
Then simply run the "main.py" script.2, 3
The model consists of five simple recommendation algorithms which are combined using a tailored strategy.
Notes:
1 The data can be downloaded from: https://ml-challenge.mercadolibre.com/downloads
2 This script requires more or less 24 GB of RAM to run.
3 Final predictions may differ a bit from the ones submitted due to the randomness of the word2vec training algorithm.