SPSS® Modeler can be used in conjunction with an IBM® SPSS Collaboration and Deployment Services repository, enabling you to manage the life cycle of data mining models and related predictive objects, and enabling these objects to be used by enterprise applications, tools, and solutions. IBM SPSS Modeler objects that can be shared in this way include streams, nodes, stream outputs, projects, and models. Objects are stored in the central repository, from where they can be shared with other applications and tracked using extended versioning, metadata, and search capabilities.
Before you can use SPSS Modeler with the repository, you need to install an adapter at the repository host. Without this adapter, you may see the following message when attempting to access repository objects from certain SPSS Modeler nodes or models:
The repository may need updating to support new node, model and output types.
For instructions on installing the adapter, see the SPSS Modeler Deployment Installation guide, available as a PDF file as part of your product download. Details of how to access IBM SPSS Modeler repository objects from IBM SPSS Deployment Manager are given in the SPSS Modeler Deployment Guide.
The following sections provide information on accessing the repository from within SPSS Modeler.
Extensive Versioning and Search Support
The repository provides comprehensive object versioning and search capabilities. For example, suppose that you create a stream and store it in the repository where it can be shared with researchers from other divisions. If you later update the stream in SPSS Modeler, you can add the updated version to the repository without overwriting the previous version. All versions remain accessible and can be searched by name, label, fields used, or other attributes. You could, for example, search for all model versions that use net revenue as an input, or all models created by a particular author. (To do this with a traditional file system, you would have to save each version under a different filename, and the relationships between versions would be unknown to the software.)
The single sign-on feature enables users to connect to the repository without having to enter username and password details each time. The user's existing local network login details provide the necessary authentication to IBM SPSS Collaboration and Deployment Services. This feature depends on the following:
- IBM SPSS Collaboration and Deployment Services must be configured to use a single sign-on provider.
- The user must be logged in to a host that is compatible with the provider.
For more information, see Connecting to the Repository.