This is the best guide for "Building recommendation systems from scratch" by Ahmed Djebali.
The notebook introduces recommendation systems and its applications in a nutshell. Then get you started and build a simple engine. And as you move along, you will learn how to build a real-time recommendation engine, a scalable recommendation engine, and more with popular frameworks such as R, Python, Spark, Neo4j, and Hadoop. You will also leran the best practices for buidling recommender systems, pros and cons of each recommendation technique and which one suits you the best.
To get started the following tools must be installed:
- Anaconda
- Jupyter notebook
- Python
- R studio
- R within Jupyter notebook
- Spark
- Node4j
- Mahout
-
The attached file is a step-by-step notebook to recommendations from zero to hero.
-
To get there just cd into the main directory (where the notebook is located), then run the notebook
$jupyter notebook. All the instructions are in there.