Low-Latency Prediction Lookup
Even the most accurate Machine Learning Pipelines are only useful if they’re driving real value for your organization. That’s why Cortex makes it easy to connect predictions directly to your business workflows in a variety of ways. In this guide, we’ll walk through one of those ways: fetching a particular prediction through Vidora’s low-latency Predictions API.
Note that this guide will serve as a general overview of the prediction lookup functionality provided by the Predictions API. Please visit the Predictions API documentation for more tactical information on how to implement the API.
What is the Predictions API?
The Predictions API provides low-latency access to predictions made by any pipeline in your Cortex account. The Predictions API returns up-to-date predictions directly to your digital products and business tools, enabling you to build automated mechanisms which act on these predictions in intelligent ways.
When should I use the Predictions API?
The Predictions API is best suited for “online” use cases where predictions are deployed on-demand and in real-time. A typical use case involves dynamic user experiences, like one of the below:
- Targeted Onsite Promotion: Dynamically display a promotional banner ad if the user has a >20% likelihood of clicking on the promoted product.
- Personalized Onsite Modules: Optimize the order and layout of content modules depending on the user’s predicted affinity to each content category.
- Dynamic Paywall: Provide a dynamic paywall experience depending on the user’s likelihood of converting to a paid subscriber.
How do I use the Predictions API?
There are two types of requests that you can make to the Predictions API.
- Get the latest prediction from a particular pipeline for a given user
- Get the latest predictions from all pipelines for a given user
To frame these requests in real terms, the above assumes a Cortex account configured to make predictions about users. Note however that your Cortex account can be configured to predict for any type of object tied to your event data (e.g. commerce items, media content, home listings, etc.).
See the Predictions API documentation for more information about how to use the API.
Still have questions? Reach out to firstname.lastname@example.org for more info!