eCommerce Recommendation Engine -

Business Goal:

To provide personalized recommendations based on consumer and socio-economic profile and purchase intelligence to increase up-selling and cross selling opportunities and to design personalized targeting and retargeting campaigns

For more information, contact us

Challenge:

Consumers have short attention spans and eCommerce engines need to have the ability to keep high levels of engagement. In addition, eCommerce stores also need to deal with the issues of cart abandonment and out of stock items, which generally drive engagement down.

Solution:

A personalized recommendation engine based on the following machine learning algorithms was developed:

Collaborative Filtering – Matrix Factorization
Semi Restricted Boltzmann Machines
General Factorization Framework – Context-aware recommendations
Alternating Least Squares (ALS) Learning

Recommendation Engine Architectural Components

Technology stack: Django, Spark, Kafka, HDFS, Gobblin, MongoDB.