What’s New in SAS Enterprise Miner 7.1 M1
Overview
SAS Enterprise Miner
7.1M1 is a maintenance release for the SAS 9.3 system. This release
updates and improves many aspects of the software and introduces a
new set of changes and enhancements. It is recommended that customers
apply the maintenance release to receive the best user experience.
The first maintenance
release of SAS Enterprise Miner 7.1 includes the following new features
and enhancements:
-
support for high-performance data
mining
-
a new incremental response node
-
enhancements to these nodes
-
Scorecard Node in Credit Scoring
-
-
-
-
-
changes to the EM migration macro
-
a new procedure for PMML scoring
Support for High-Performance Data Mining
SAS has invested in
developing a key set of statistical and data mining tasks that execute
on a dedicated high performance appliance. The software will distribute
data, memory, and computations over a grid of systems producing dramatic
improvements in large data scalability and run times. SAS Enterprise
Miner 7.1M1 uses the SAS High Performance system for building predictive
models. New procedures cover data binning, imputation, sampling,
decisions, logistic and linear regressions, neural networks, random
forests, scoring and transformations. New nodes integrate those functions
into the Enterprise Miner environment. Contact your SAS sales team
for more information.
New Incremental Response Node
A new Incremental Response
Node introduces modeling the marginal effect of a treatment on individual
outcomes. The model answers questions such as which customers are
more likely to purchase only due to a new sales campaign, rather than
customers who would have purchased without the campaign. These customers
are known as true responders. This new tool is experimental in this
release.
Enhanced Nodes
Scorecard Node in Credit Scoring
The Scorecard Node adds
the Accuracy Ratio to fit statistics and the corresponding Accuracy
Profile Chart.
Interactive Grouping Node
The user interface for
the Interactive Grouping Node has been redesigned. The coarse detail
and fine detail tabs have been consolidated to a single tab. This
consolidation allows users to see and edit the relevant information
without having to toggle between tabs.
Ratemaking Node
A new property can be
used to set the reference level for each variable and report the 95%
Wald Confidence Limits in the Parameter Estimates table. The Relativity
Plots include a new line band plot in which the upper and lower bound
confidence intervals are displayed.
Survival Node
Users can now choose
to enter the cubic spline basis functions as part of the stepwise
variable selection procedure in addition to the main effects.
Multiplot Node
This node has been modified
to provide an auto scroll function for plots. The node will automatically
create plots for all input and target variables in your data set.
The new control helps users cycle through the plots automatically
to search for interesting patterns and relationships.
EM Migration Macro
The behavior of the
project migration macros has changed to make processing project data
views optional. This avoids some situations that can display errors
in the log that are not important to the project file migration process.
PMML Scoring
A new procedure is available
for scoring data based on models saved as PMML documents. PMML is
a multi-vendor standard for encoding data mining models and is most
used in scoring processes. This feature is experimental in this release.
Contact SAS Technical Support for more information.
Copyright © SAS Institute Inc. All rights reserved.