Add a Model via Web Browser Client

This procedure describes how to add a new model in to the Analytics Scoring Repository via the Analytics Scoring Repository web browser client in order to make the model available for deployment in a Spectrum™ Technology Platform dataflow.

Using the View Models procedure, navigate to the Repository page.
Click the Publish a model button (located directly above the model table) to navigate to the Add new model page.
The Publish a model button will be disabled if any models are selected in the current page of the model table. Unselect these models to enable the Publish a model button.
On the Add new model page there are three input fields: Name, Description, and Upload model. Write the name of the new model in to the text box labeled Name.
Model names must be unique. If a model name is already in use by another non-deleted model stored in the Analytics Scoring Repository, the text box will be highlighted with a validation error.
Note: White space at the start or end of a model name is always removed. This means if the name entered is already in use, adding white space to the start or end of the name will not resolve the error.

Note: Model names are case sensitive. This means models can share the same name so long as they have different uppercase and lowercase characters. For example models named "my Model", "my model" and "mY ModEl" can co-exist simultaneously within the repository.

Note: Name validation only detects conflicts against the contents of the Analytics Scoring Repository when the page was opened. This means if the repository has changed since navigating to the Add new model page, for example, models have been added or deleted by other users, then the page validation will not have these changes. However, after you click the Publish button, the new model will be re-validated against the up-to date version of the Analytics Scoring Repository .
Write a description for the new model in the Description box.
This field is required. Models must have a description.
Add a model definition by clicking the Choose file button and selecting a model stored in either QMML or PMML file formats.

Note: The selected model must be less than 50MB in size.

Note: When using an older browser that does not support file picking, the XML model definition must be input into the text area labeled Upload model.

Note: If the selected model is a QMML model, the QMML must have at least one compiled live ruleset for the model to be successfully published. In the case where there are multiple compiled live rulesets the first such ruleset found within the QMML will be used for the model definition.

Note: If you wish to compile and/or change the live status of rulesets within your QMML (or indeed add or delete rulesets), you can use the Spectrum Miner qsqmmledit command-line utility within your Spectrum Miner installation. (Help for the utility can be displayed in your command-line window by typing "qsqmmledit.exe -help").

Note: The model definition will not be validated until clicking the Publish button.
Click the Publish button to begin uploading the model to the Analytics Scoring Repository .
The Publish button will be disabled until the new model has a valid name and a description and model definition specified.
Wait while the busy icon is shown on the Publish button. No changes can be made to the new model’s name, description or model definition while the model is being published. Once the model has been successfully published to the Analytics Scoring Repository, the page will navigate to the Details page for the new model.
The Cancel button is also disabled while the model is being published.

Note: If a problem is found while publishing the model, for example, the XML model definition is invalid or the model name is already in use, then the model will not be published and an alert error will be displayed detailing the problem. Click OK to close the error alert and correct the problem before attempting to publish the model again.