yadavak/mulit-criteria-collaborative-filtering-recommendation-system ? reverse-engineered prompt

Build me the multi criteria collaborative filtering recommendation system from the paper “A novel deep multi criteria collaborative filtering model for recommendation system” by Nour Nassar, Assef Jafar, and Yasser Rahhal.

I want this repo turned into a working Python project that can train on a ratings style dataset, handle multiple criteria for each item, and produce recommendations for users. Please make the code clean and runnable from the README, with a simple way to install dependencies and run training and evaluation. If the paper leaves any parts unclear, look up the paper or current docs online and make reasonable choices so the implementation is complete and practical.

Please also add a small example or demo script so I can see it working end to end, plus basic documentation on how to prepare data, train the model, and get recommendations. If there are any missing pieces in the existing files, fill them in instead of leaving the project half done.