In this scoring experiment, we create a web service that predicts for new flights whether scheduled passenger flight is delayed or not.
This is a scoring experiment created from the "Binary Classification: Flight delay prediction" sample experiment which is described in detail here: https://gallery.azureml.net/Details/837e2095ce784f1ba5ac623a60232027 In this scoring experiment, a web service is created that predicts for new flights whether or not they will be delayed. This experiment was created by highlighting the Boosted Decision Tree module of the training experiment, then selecting the button "Create scoring experiment." This makes it so the trained model from the training experiment is applied in the scoring experiment (without the need to train the model again - in scoring experiment/web service, it is using the trained model to simply score new points). Here, I chose to have the input be the processed data that is from both the flight and weather data combined (chose to "save as dataset" after the last data processing step and used that data in the experiment, deleting the other modules from the experiment). Created by a Microsoft employee.