Add a Model via Spectrum Miner Spectrum Connector

The Spectrum Connector integration component enables the publishing of QMML models into the Spectrum Analytics Scoring Repository directly from Spectrum Miner.

The Spectrum Connector integration component can be installed as an addition to any existing Spectrum Miner version 7.0B (or above) installation, including 'desktop' installs.

There are three different methods of publishing models to the Analytics Scoring Repository using the Spectrum Connector integration component:

Publishing QMML Models from Spectrum Miner

Publishing QMML Models from Decision Studio Model Builder tools

Publishing a QMML model from the Spectrum Miner command line tool

Publishing QMML Models from Spectrum Miner

This procedure describes how to add a new model in to the Analytics Scoring Repository using Spectrum Miner.

Open Spectrum Miner and in the File Browser tab navigate to the directory containing the .qmml file you wish to publish.
Right click on the .qmml file and select 'Publish -> to Spectrum Analytics Scoring Repository '.
When prompted enter a valid Spectrum username and password to login into the Spectrum platform. The login prompt will be displayed once per Spectrum Miner session with any subsequent model publishing using the credentials previously provided.
In the model details tab specify the model name, model description and QMML ruleset for the model to be published into the Spectrum Analytics Scoring Repository .
The name given to the model must be unique within the Spectrum Analytics Scoring Repository.
Click the 'Go' button.

Publishing QMML Models from Decision Studio Model Builder tools

This procedure describes how to add a new model in to the Analytics Scoring Repository directly from one of the Decision House model building tools.

In the chosen Decision House Model Builder tool, build your model.
Press the 'Publish [type of model] to Spectrum' button that was enabled after the model build completed in the previous step.
When prompted enter a valid Spectrum username and password to login into the Spectrum platform. The login prompt will be displayed once per Spectrum Miner session with any subsequent model publishing using the credentials previously provided.
In the model details tab specify the model name, model description and the output fields for the model to be published into the Spectrum Analytics Scoring Repository .
The name given to the model must be unique within the Spectrum Analytics Scoring Repository.
Note: Only fields with their associated checkbox selected will be included as output fields in the model. At least one field must be selected.
Click the 'Publish' button.

Publishing a QMML model from the Spectrum Miner command line tool

This procedure describes how to add a new model in to the Analytics Scoring Repository using the spectrumpublish data build command line tool provided by Spectrum Miner's Spectrum Connector integration component.

At a command prompt, execute the spectrumpublish utility e.g. '$QSHOME/server/qs7.0B/win32/bin/spectrumpublish -input [qmml file path] -name [model name] -description [model description] -ruleset [ruleset name] -username [Spectrum username] -password [password]

Note: The -input, -name, -username, -password and -description arguments must be specified. The -ruleset argument is optional.

Note: In the case where the supplied QMML file contains multiple rulesets but the -ruleset argument is not specified, then only the first ruleset within the QMML document is published into the Spectrum Analytics Scoring Repository .