Welcome to InterSystems IRIS Cloud IntegratedML!
The IntegratedML featureOpens in a new tab of InterSystems IRIS® Data Platform enables organizations without extensive machine learning expertise to define and execute predictive models by applying automated functions directly from SQL. The InterSystems IRIS Cloud IntegratedML service, available as an optional feature of the Cloud SQL service, gets you started on machine learning without the burden of provisioning, configuring, and maintaining cloud infrastructure.
Using Cloud IntegratedML
To get started with InterSystems IRIS Cloud IntegratedML, load the InterSystems Cloud Services PortalOpens in a new tab to sign up for a Cloud SQL trial or subscription and create a deployment with IntegratedML enabled.
Next, open your deployment in the Portal to create and load the tables you need for the machine learning process by importing DDL, DML and CSV data files. (The Portal also lets you review and manage system information, service requests, users and tenants, and performance profiles.)
When you are ready, go to the IntegratedML Tools page in the Portal, which guides you through the four stages of the machine learning process:
-
Create a modelOpens in a new tab, which defines the elements of a predictive model, including the input fields (features), the predicted field (label), and the data types of these fields.
-
Train the modelOpens in a new tab using the configured provider, which uses a structured process to compare the performance of different machine learning model types (linear regression, random forest, and so on) with the data.
-
Validate the trained modelOpens in a new tab against data like the data it was trained on, returning simple metrics for both regression (data range) models and classification (category) models.
-
PredictOpens in a new tab the output field (label) value for each row in a set of data containing the input fields (features) the model was trained on. Classification models can also call on the provider to return the probability that the predicted value is the correct result for the model.
You may want to start with the IntegratedML Orientation that is available throughout the Portal.
You can also use IRIS Cloud IntegratedML by connecting to it from your Python, Java, .NET, or C++ application in just three easy steps:
-
Download the InterSystems driverOpens in a new tab for DB-API , JDBC, ADO.NET, or ODBC.
-
Get the connection information for your deployment from the Deployment Details section of the Overview page in the Portal, and confirm that external connections are enabled.
-
Add the connection string to your application as described in Connecting Your Application to InterSystems IRISOpens in a new tab.
Your application can then make full use of the deployment's SQL and IntegratedML functionality (although you must use the portal to upload or transfer any needed data files to the deployment).
For help in addressing any problems you may encounter, either in executing IntergratedML SQL statements or in connecting programmatically, see the troubleshooting guideOpens in a new tab.
Learning Resources
To learn more about the machine learning and IntegratedML, explore the following resources:
-
Documentation: Using IntegratedMLOpens in a new tab (Bear in mind that this documentation covers the general use of IntegratedML, not the use of the Cloud IntegratedML service in particular.)
-
Videos: What is IntegratedML?Opens in a new tab and Building Models with IntegratedML in the CloudOpens in a new tab
To learn more about InterSystems SQL, see Learning Resources in Welcome to InterSystems IRIS Cloud SQL.