This project intends to reproduce Translation-based Recommendation using Python, though, the authors published official C++ code.
Python 2.7
-
Clone the entire project
-
You could of course download raw dataset to root directory from http://jmcauley.ucsd.edu/data/amazon/links.html, then execute Datapreprocessing.py and DataPartition.py in order to get more structured dataset packaged in numpy format.
-
For your convenience, alternatively, you can basically run "src/TransRec.py" and other baselines, e.g. "FPMC.py", since numpy datasets of a few categories already exist.
-
To change different dataset category, e.g. "Automotive", for training and evaluation, put your category name here.
dataset_name = 'Automotive'