Loan ChargeOff Prediction

Operationalization with ML Server


To use ML Server Operationalization services from your local computer, you must first connect to the edge node using the steps below.

Connect to Edge Node

  • Windows users: For instructions on how to use PuTTY to connect to your HDInsight Spark cluster, visit the Azure documentation. Your edge node address is of the form CLUSTERNAME-ed-ssh.azurehdinsight.net.
  • Linux, Unix, and OS X users: For instructions on how to use the terminal to connect to your HDInsight Spark cluster, visit this Azure documentation. The edge node address is of the form sshuser@CLUSTERNAME-ed-ssh.azurehdinsight.net
  • All platforms: Your login name and password are the ones you created when you deployed this solution using the 'Deploy to Azure' button on the Quick start page

Configure Deployment Server

  • Once you have connected to the edge node you can access the Administration Utilities for the web server with:
sudo dotnet /usr/lib64/microsoft-r/rserver/o16n/9.1.0/Microsoft.RServer.Utils.AdminUtil/Microsoft.RServer.Utils.AdminUtil.dll

Your server has been configured with username admin and HDInsight Cluster Login Password (the password that you have provided during deployment). You can use this utitlity to change the password if you wish. (If you do so, you will need to change the password in the loanchargeoff_deployment.R and loanchargeoff_web_scoring.R script.)

You can also use this utility to check on the status of the web server.

  • Enter 6 to select “6. Run diagnostic tests”;
  • Enter a to select “A. Test configuration”;
  • Provide username as admin and the password you just created;
  • You should see “Overall Health: pass”;
  • Now press e followed by ‘8’ to exit this tool
Do not close the terminal window - it should remain open when you execute loanchargeoff_web_scoring.R to try the web API on your local computer, after it is created with loanchargeoff_deployment.R from RStudio the cluster.

Return to Typical Workflow