Building Personalized Search Experiences with Cortex

Personalized Search involves surfacing results that are tailored to the unique preferences of the user making a search query. With personalized search, you can maximize engagement. This happens because you can expose relevant content that the user is predisposed towards.

Enabling Personalized Search with Cortex

Cortex Automation allows our partners to build machine learning technology directly into their digital products. This therefore enables intelligent and automated consumer experiences. Those include experiences such as Personalized Search. Cortex Automation creates a closed-loop system through real-time APIs. As a result, this system constantly ingests raw behavioral data. It then returns up-to-date predictions in 10-20ms.

Vidora’s User Items Ranking API endpoint powers Personalized Search.

  1. User makes a search query on your site
  2. You (or a third party) returns a list of results relevant to the search query
  3. A call is made to VIDORA’s User Items Ranking API with the following parameters
    1. ID of the user
    2. List of IDs of the search query result
  4. Vidora’s User Items Ranking API returns the list of query results in the order of preference for the specific user

 

How Does It Work?

The User Items Ranking API leverage Cortex’s Recommendation models, which use a machine learning technique known as collaborative filtering to determine each users’ unique preferences. Cortex then generates personalized recommendations. Those recommendations use a weighted combination of the following three signals.

  1. Personalized preference – what other items has the user interact with in the past?
  2. Popularity trends – is the item globally popular among all users, and increasing in popularity over time?
  3. Item Recency – has the item been recently released?

After measuring these three signals, Cortex trains on past data to determine how to weight each signal in order to maximize engagement with its recommendations.

About Vidora

Vidora enables anyone in any business to build and use complex machine learning models. With Vidora’s self-service platform, Cortex, machine learning is intuitive, interpretable and fast. Cortex also automates the entire machine learning pipeline from raw data to model outputs. Experts in machine learning and artificial intelligence from Stanford, Berkeley and Caltech developed Cortex. Finally, Cortex sits at the heart of some of the largest global brands, such as Walmart, News Corp, and Discovery.

Posted by Michael Firn on November 16, 2018