Explore Healthcare Use Cases

Explore three example use cases of healthcare machine learning model training and deployment by uploading Jupyter Notebook and using it directly with Oracle Cloud Infrastructure Data Science service.

Provision OCI Data Science Service

Provision Oracle Cloud Infrastructure Data Science using Oracle Cloud Infrastructure Resource Manager and Terraform.

1. Go to GitHub.
2. Clone or download the repository to your local computer.
3. Review and follow the instructions in the Readme.
4. Assign your user to the group DataScienceGroup to access the service.

Use the Example Notebooks in OCI Data Science Service

Create a Jupyter Notebook session in Oracle Cloud Infrastructure (OCI) to explore Oracle Cloud Infrastructure Data Science service healthcare models. The repository in GitHub provides demos, tutorials, and code examples that highlight various features of OCI Data Science service and AI services.

The following machine learning healthcare models are available for download:

• Predict health of fetuses based on cardiotocogram signals.
• Predict Parkinson disease from variations in speech characteristics.
• Predict breast cancer from biospy cell image characteristics.

1. Log into the OCI Console.
2. Navigate to Analytics & AI, click Data Science, then in the compartment of your choice, click Create Project.
3. Select the project, and then click Create Notebook Session.
4. Follow the prompts to define a name and the shape of the machine to use for this session. The default block storage is sufficient.
5. Click the Open button to launch the OCI Data Science Notebook session.
   The launcher opens as the default page in the Jupyter Notebook interface.
6. Scroll down and click the Terminal icon to launch a new terminal window.
7. Install a general purpose machine learning conda environment with the command:

   odsc conda install -s generalml_p37_cpu_v1

8. Download the sample notebooks from GitHub.

   wget https://raw.githubusercontent.com/oracle-samples/oci-data-science-ai-samples/master/labs/healthcare-use-cases/Breast_Cancer.ipynb
   wget https://raw.githubusercontent.com/oracle-samples/oci-data-science-ai-samples/master/labs/healthcare-use-cases/CTG_fetal_health.ipynb
   wget https://raw.githubusercontent.com/oracle-samples/oci-data-science-ai-samples/master/labs/healthcare-use-cases/Parkinson.ipynb

9. On the file browser pane (on the left), double-click on a notebook to explore.
10. On startup, select the kernel generalml_p37_cpu_v1 kernel.
11. Browse the notebook by running each cell using the icon on the top menu bar.

    Each notebook goes over some data exploration steps, data visualization of the various features, data transformation to prepare for model training, and training on various models to estimate the best algorithm. Once a suitable model is selected and trained, it is stored into the Model Catalog and deployed as a Model Deployment.