IBM 15 - page 19

New Features

SQLgeneration e nhancements. The Aggregate no de now supports SQL generation for date, time,

timestamp, and string data types, in addition to integer and real. With IBM Netezza databases, the

Sample node supports SQL generation for simple and complex sampling , and the Binning node

supports SQL generation for all binning methods except Tiles.

In-databasem odel scoring. For IBM DB 2 for z/OS, IBM Netezza and Teradata databases, it is

possible to enable SQL pushback of many of the model nuggets to carry out model scoring (as

opposed to in-database mining) within the database. Todo this, you c an install a scoring adapter

into the database. When you publish a model for the scoring a dapter,the mo del is enabled to use

the user-deﬁned function (UDF) capabilities of the database to perform the scoring.

A new conﬁguration option, db_udf_enabled in options.cfg, causes the SQL generation option

to generate UDF SQL by default.

Newformat for database connection in batc h mode. The format for specifyin ga dat abase connection

in batch mode has changed to a single argument, to be consistent with the way it is speciﬁed in

scripting.

Enhancementsto SPSS Sta tisticsin tegration. On the Statistics Output node, add itional procedures

are available on the Syntax tab through the

Select a dialog button. The Regression s ubmenu

now supports Partial Least Squares regression, and there is a new Forecasting s ubmenu with the

following options: Spectral Analysis, Sequence Chart s, Autocorrelations, and Cross-correlations.

For more information, see the SPSS Statistics documentation.

The Syntax tab of the Statistics Output node also has a new option to generate a Statistics File

source node for importing the data that results from running a stream containing the node. This is

useful where a procedure writes ﬁelds such as scores to the active dataset in addition to displaying

output, as these ﬁelds would otherwise not be visible.

Non-rootuser on U NIX servers. If you have SPSS Modeler Se rver installed on a UNIX server, you

can now install, conﬁgure, and start and stop SPSS Modeler Server as a non-root user without the

need for a private password database.

Deployed streams can now access IBM SPSS Collaboration and Deployment Services model

managementfe atures. When a stream is deployed to IBM SPS SC ollaboration and Deployment

Services as a stream, it can now use the same model management features as it could if deployed

as a scenario. These features include evaluation, re fresh, score, and champion/challenger.

Improvedmethod of chang ingO DBC connection for SPSS Modeler stream and scenario job steps. For

stream and scenario job steps in IBM SPSS Collaboration and Deployment Services, changes to

an ODBC connection and related logon credentials apply to all related job steps. This means that

you no longer have to change the job steps one by one.

Choiceof execu tion branch in deployed streams. For st ream job steps in IBM SPSS Collaboration

and Deployment Services, if the stream contains branches you can now choose one or m ore

stream branches to execute.