In this hands-on lab, you are working as a data scientist for AIDoctor, a medical clinic. You’ve been asked to come up with a model that can be used to predict if a patient is diabetic based on factors such as age, blood pressure, BMI, and more. You’ll use Azure Machine Learning studio and Azure Machine Learning designer to create a classification model that will predict if a patient is diabetic or not. You’ll also get to create an inference pipeline and deploy the model.
Successfully complete this lab by achieving the following learning objectives:
- Open the Azure ML Studio and Create a New Dataset
In the Azure ML studio, create a new dataset using the diabetes data file from Microsoft.
- Review the Dataset
Take a look at the data you’re working with and see if there are any missing values or values that will cause issues.
- Design a Pipeline to Predict if Patients Have Diabetes and Review Evaluation Results
Drag and drop modules to create a pipeline that will train a classification model. After it runs, review the evaluation results and metrics.
- Create and Run an Inference Pipeline
Create an inference pipeline, change the Web Service Input module’s output to the Apply Transformation module, and remove the Diabetic column in the Select Columns in Dataset module. After those modifications, run the pipeline.
- Deploy the Model to Azure Container Instance and Test It
Deploy the model to an Azure Container Instance, and test the endpoint.
Note: Make sure you click Advanced before deployment and change both the CPU reserve capacity and memory reserve capacity to 1 or the resources will be deleted.