Network Based Recommendation Engine using Map-Reduce (in Python) to run on top of Hadoop!

The recommendation engine is divided into three parts :

1. Pre-process Job
2. Resource Allocation Job
3. Recommendation Job

Each job has their own mapper and reducer.

MongoDB is used in this version for real-time data storage and streaming.

Algorithm explanation link

Movie-Lens data set : ml-1m is used as reference!

To run python map-reduce programs we need Hadoop Streaming. Command for hadoop streaming is inside bin directory of this project, please replace directories as per your need.