Recommendation Systems

Recommendation Systems

Recommendation systems are essential part of every e-commerce business today. It helps the business by making sure that customers have a good experience in finding out products and finally making a purchase. There are mainly two business goals that a recsys(I will refer to Recommendation systems by the acronym recsys now onwards) helps improve.

1. Improving the conversion of existing customers

2. Improving Retention

Recsys have been studied and researched for a long time. As a result, there are many models/algorithms which have been invented to tackle recommendations. Recsys can be broadly divided in following categories.

1. Non-personalized or Content based recommendation

2. item-item collaborative filtering and user-user collaborative filtering(Nearest neighbor recommendor)

3. matrix factorization based collaborative filtering

4. Hybrid recommendor systems

All existing techniques of recsys can be mapped to one of these categories. We are going to look at each of these in separate blog posts.

I would list down some good places where you can find more material on recsys.

1. [recsys specialization on coursera] (https://www.coursera.org/specializations/recommender-systems)

2. [microsoft book on recsys] (http://mbmlbook.com/Recommender.html)

3. [google on recsys] (https://developers.google.com/machine-learning/recommendation)