online-cf

Real-time collaborative filtering for item-item similarities

An implementation of the paper Cloud based real-time collaborative filtering for item-item recommendations.

We also stored which items every user rated and hence improved the performance.

The cosine branch has an experimental implementation of similarity calculation using numeric ratings instead of binary data.

Dependencies

You must have redis.py. Install it with this instruction:

$ pip3 install redis

Instructions

Scripts are written in Python 3.

Run the following redis servers running before running the scripts:

# cd redis.../src

$ redis-server --port 6379
$ redis-server --port 6380
$ redis-server --port 6381
$ redis-server --port 6382

Run the following python scripts via run script:

$ chmod +x run
$ ./run

Feed the scripts with data:

$ python3 client.py

Monitor the most similar items to an item live with similars.py script:

$ python3 similars.py 1
Showing the most similar items for 1
[['1220', '0.036'], ['315', '0.036'], ['4886', '0.036'], ['780', '0.036'], ['1923', '0.032'], ['586', '0.032'], ['4973', '0.030'], ['1198', '0.021'], ['457', '0.017'], ['296', '0.014']]
.
.
.