How is scoring done in azure ml?

Question

Suppose I have 50 features in my dataset, but after feature engineering(one hot encoding or tf-idf) I get 200 feature colums. The model is trained on these 200 features and is deployed and now we have a rest endpoint for the model. Now the customer will hit the endpoint with 50 features but the model is expecting 200 columns. Where will the conversion of 50 to 200 features takes place?

Answer 1

Hi, thanks for reaching out. It seems you're wondering how to consume the model after feature engineering. The simple answer is that you'll need to apply the same transformations on your testing dataset. The schema of the input dataset should match the schema of the data used to train the model.

---

--please don't forget to Accept Answer if the reply is helpful. Thanks.--