From the course: MLOps with Databricks
Join today to access over 25,500 courses taught by industry experts.
Feature table and feature function - Databricks Tutorial
From the course: MLOps with Databricks
Feature table and feature function
- [Instructor] In the previous chapter, we trained, logged, and registered a sklearn pipeline. We assume that all the features required for the model should be passed to the predict function to generate the prediction. What if we need to look up and even pre-compute some features for predictions? Let's go back to our hotel booking continuations dataset. There are some historical features, like the number of previous cancellations, and the number of the previous non-canceled bookings, and whether the booking is repeated. Those can be looked up in the feature table. We can calculate the lead time, based on the reservation and arrival date. To accomplish that, we need to train a model using a training set class. To construct a training set, we need a DataFrame, a Feature table, and the Feature function. This is what we focus on in this video. In the next video, I will show how to create a training set and use the training model. The code that uses feature engineering package may not work…
Download courses and learn on the go
Watch courses on your mobile device without an internet connection. Download courses using your iOS or Android LinkedIn Learning app.