Machine Learning Pipeline Automation

Loading...
The Machine Learning Pipeline Automation (MLPA) API enables CRUD operations on automation projects, which automates VDMML project creation, pipeline building and training, and the production of champion models.

AutomationProjects

Contains the operations for working with machine learning pipeline automation.
get
Return all automation projects
Internal-Use OnlyThe endpoint to get all available automation projects.
/projects
post
Create an automation project
Internal-Use OnlyA user begins using the service by creating an automation project. The Machine Learning Pipeline Automation API uses the following attributes: user-supplied project name, data table URI and target variable, and other attributes. The service creates an underlying analytics (VDMML) project, executes multiple steps of pre-processing, builds a data analysis pipeline, and runs the pipeline to completion. The entire process is fully automated without further user intervention. The request body is structured in three sections.
/projects
head
Get headers for an automation project
Internal-Use OnlyThe endpoint to get headers for a specific automation project specified by an ID.
/projects/{projectId}
get
Get an automation project
Internal-Use OnlyThe endpoint to get a specific automation project specified by an ID.
/projects/{projectId}
put
Update the specified automation project
Internal-Use OnlyIn some cases, the user could select the wrong data set or target variables accidentally when creating an automation project. The project might fail to run or the result might not be what the user intended. The ability to update an automation project comes to rescue in those cases, where the user can update the project's attributes and settings with the following request and rerun the project. This API is solely used for updating a project’s parameters. There is a separate endpoint for updating the state of a project, for example, stopping or restarting a project run. We do not support PATCH operation to update the automation project with changes only. According to standard, the user must enter the full project info in this PUT request, whether or not the parameters are to be changed. List of parameters that can be updated.
/projects/{projectId}
delete
Delete an automation project
Internal-Use OnlyWhen it is not needed anymore, the automation project can be deleted with the request below. By default, this API deletes the automation project while keeping its associated analytics project. To delete both projects, set the query parameter "propagate" to true.
/projects/{projectId}
head
Retrieve headers for completed project reports
Internal-Use OnlyRetrieves headers for a list of reports for a project that has completed modeling.
/projects/{projectId}/reports
get
Retrieve reports for a completed project
Internal-Use OnlyRetrieves a list of reports for a project that has completed modeling.
/projects/{projectId}/reports
put
Retrain the specified automation project
Internal-Use OnlyTo retrain an automation project with changed parameters, use the retrainProject endpoint. By default, the service generates a new automated pipeline. This behavior can be overwritten with an optional query parameter replacePreviousPipelines. When set to true, the parameter instructs the service to remove all previous automatically generated pipelines before creating a new pipeline.
/projects/{projectId}
head
Get the headers for champion model information for a completed automation project
Internal-Use OnlyGet the headers for champion model information for a completed automation project specified by project ID.
/projects/{projectId}/models/@championModel
get
Get the champion model information for a completed automation project
Internal-Use OnlyGet the champion model information for a completed automation project specified by project ID.
/projects/{projectId}/models/@championModel
put
Register or publish the champion model of the automation project
Internal-Use OnlyRegister or publish the automation project's champion model.
/projects/{projectId}/models/@championModel
head
Get the headers for champion model reports for a completed automation project
Internal-Use OnlyGet the headers for champion model reports for a completed automation project specified by project ID.
/projects/{projectId}/models/@championModel/reports
get
Get the champion model reports for a completed automation project
Internal-Use OnlyGet the champion model reports for a completed automation project specified by project ID. If the reportName parameter is not included, a list of default reports are returned. The reportName parameter can be used to return a sub set of these reports. The default reports returned for a binary target are iterplot, varimportance, dmcas_lift, dmcas_roc, dmcas_fitstat, and dmcas_misc. The default reports returned for an interval target are iterplot, varimportance, dmcas_lift, and dmcas_fitstat. The default reports for a nominal target are dmcas_lift, dmcas_roc, dmcas_fitstat, dmcas_misc, dmcas_miscTable and dmcas_nommisc.
/projects/{projectId}/models/@championModel/reports
post
Score data with the champion model of the automation project
Internal-Use OnlyScore data with the automation project's champion model.
/projects/{projectId}/models/@championModel/scoreData
head
Get headers for current state of an automation project
Internal-Use OnlyGet headers for the request that gets the current state of an automation project
/projects/{projectId}/state
get
Get current state of an automation project
Internal-Use OnlyAutomation project state can be one of these enum values.
/projects/{projectId}/state
put
Update state of the specified automation project
Internal-Use OnlyUpdate state of the automation project with the ID from the path. The automation project state is considered a single piece of information and operations on state are atomic. State update is completely controlled by the state machine based on current state and intended state.
/projects/{projectId}/state