Azure Machine Learning Prediction - Input and Outputs

Question

I am attempting to follow this tutorial however I was attempting to predict MPG for a set of cars rather than oil prices and have the following set up:

1. MPG Sample dataset
2. Remove missing values, project everything (weight, displacement, cylinders, etc) except model name
3. Split 75 to train model, 25 to score model
4. Train model on MPG column with neural network
5. Score model which is fed by Train Model and Split
6. Score model is fed to Evaluate model

This all seems to run fine and without issue, so I create a scoring experiment and then publish it as a web service, however when I attempt to input values it is asking for an MPG input. My understanding is that this would be the predicted value, so it seems somewhat opposite to have to enter this as a value, or am I just understanding a basic tenet of machine learning?

In short: Ideally I would like to be able to enter everything but the MPG and get a prediction on what the MPG is for a given set of value.

Answer 1

You could also add project columns to exclude label as part of scoring experiment and connect web service output port to the output of project columns