Diagnose Cancer with an Amazon Machine Learning Classifier


52 students completed the lab in ~36m

Total available time: 1h:0m

24 students rated this lab!

Lab Overview

Binary classification is a kind of machine learning that predicts whether an item belongs to one of two classes. Example applications of binary classification are predicting if an email is spam or not, if a user will buy a new product, or determining if a tissue sample is benign or malignant. In this Lab, you will learn about binary classification and model evaluation in AWS as you diagnose cancer with an Amazon Machine Learning binary classifier. You will train a model with medical data, evaluate the model's performance, and use the model to make diagnoses.

Lab Objectives

Upon completion of this Lab you will be able to:

  • Create Amazon Machine Learning datasources, and binary classification models
  • Understand model how recipes can be used to transform data
  • Interpret model evaluation results and how model parameters impact them
  • Perform real-time predictions with the model

Lab Prerequisites

You should be familiar with:

  • Basic Amazon S3 concepts

Lab Environment

Before completing the Lab instructions, the environment will look as follows:

After completing the Lab instructions, the environment should look similar to:

Follow these steps to learn by building helpful cloud resources

Logging in to the Amazon Web Services Console

Your first step to start the Lab experience

Understanding the Diagnostic Cancer Data

Learn about the medical data you will use in the Lab

Creating an Amazon Machine Learning Datasource

Create a datasource from the medical data in the S3 bucket

Creating an Amazon Machine Learning Model

Create a model to train with the datasource you created

Inspecting the Amazon Machine Learning Datasource

Gain insights into the data by looking at datasource statistics

Evaluating the Amazon Machine Learning Model

Understand the model's performance with an evaluation

Creating Diagnoses with Amazon Machine Learning Model Predictions

Use sample measurements and the model to predict an outcome