Recommendation Engine

Most organizations want to use their wealth of customer interest data to offer more relevant products to their customers in order to increase sales or engagement. This can be directly by pushing the product on the web page or over email or through email.


Using a customer’s purchase/view/listen history, recommend additional products the customer is likely to be interested in.


Additional products this customer is likely to be interested in.


While the number of data sources is likely to be small, relying primarily on the transactional data to identify all content/products the customer has used/purchased, it can be daunting to consolidate this data into a single row of data for each customer. This can involve processing massive amounts of data and generating a massive number of features. Depending on the technique used, additional demographic data may need to be added. All this data needs to be cleaned, joined and transformed into valuable ML features before going into model training. This pre-modeling prep process can be frustrating and time consuming. We are here to help.

Modeling techniques and libraries

Collaborative filtering

Compare customers to other customers who have also bought/used similar products/content. From those neighboring customers who have a lot of items in common, suggest other items that these neighbors have in common, but that the original customer has not bought/used.

  • Sklearn
  • Scipy
  • Scikit-surprise
  • implicit

Content-based filtering

Use the attributes of the products a customer has used/purchased to find other products that have similar attributes.

  • Sklearn

Deep learning-based approach

Perform either Collaborative or content-based filtering using a deep learning framework.

  • Keras
  • Tensorflow

Related accelerators

No items found.

No-code/low-code data prep and visualization