SAS^® Enterprise Miner^™ Features

Intuitive interfaces

• Interactive GUI for building process flow diagrams.
• Batch processing code for scheduling large modeling and scoring jobs.

Data preparation, summarization & exploration

• Access and integrate structured and unstructured data sources.
• Outlier filtering.
• Data sampling.
• Data partitioning.
• File import.
• Merge and append tools.
• Univariate statistics and plots.
• Bivariate statistics and plots.
• Batch and interactive plots.
• Segment profile plots.
• Easy-to-use Graphics Explorer wizard and Graph Explore node.
• Interactively linked plots and tables.
• Data transformations.
• Time series data preparation and analysis.
• Interactive variable binning.
• Rules Builder node for creating ad hoc data-driven rules and policies.
• Data replacement.

Advanced predictive & descriptive modeling

• Clustering and self-organizing maps.
• Market basket analysis.
• Sequence and web path analysis.
• Link analysis.
• Dimension reduction techniques:
  
  • Variable selection.
  • LARS (Least Angle Regression) variable selection.
  • Principal components.
  • Variable clustering.
  • Time series mining.
• Linear and logistic regression.
• Decision trees.
• Gradient boosting.
• Neural networks.
• Partial least squares regression.
• Two-stage modeling.
• Memory-based reasoning.
• Model ensembles, including bagging and boosting.
• Time series data mining.
• Survival analysis.
• Ratemaking for insurance.
• Incremental response/net lift models.

Open-source R integration node

• Write code in the R language inside of SAS Enterprise Miner.
• Makes SAS Enterprise Miner data and metadata available to your R code and returns R results to SAS Enterprise Miner.
• Training and scoring for supervised and unsupervised R models.
• Allows for data transformation and data explorations of R models in SAS Enterprise Miner.
• Generates model comparisons and SAS score code for supported models.

Select set of high-performance procedures & nodes

• Multithreaded, high-performance procedures:
  
  • High-performance variable reduction.
  • High-performance neural networks.
  • High-performance random forests.
  • High-performance 4score.
  • High-performance decide.
  • High-performance data mining database.
  • High-performance sampling.
  • High-performance data summarization.
  • High-performance imputation.
  • High-performance binning.
  • High-performance correlation.
  • High-performance Bayesian network.
  • High-performance clustering.
  • High-performance Support Vector Machine.
• Multithreaded, high-performance nodes:
  • HP Data Partition.
  • HP Explore.
  • HP Transform.
  • HP Variable Selection.
  • HP Regression.
  • HP Neural.
  • HP Forest.
  • HP Impute.
  • HP Tree.
  • HP GLM.
  • HP Principal Components.
  • HP Cluster.
  • HP SVM.

Fast, easy & self-sufficient way for business users to generate models

• SAS Rapid Predictive Modeler automatically generates predictive models for a variety of business problems.
• Business analysts and subject-matter experts work from SAS Enterprise Guide or the SAS Add-In for Microsoft Office (Excel only).
• Models can be opened, augmented and modified in SAS Enterprise Miner.
• Produces concise reports, including variable importance charts, lift charts, ROC charts and model scorecards, for easy consumption and review.
• Ability to score the training data with an option to save the scored data set.

Model comparisons, reporting & management

• Assessment features for comparing multiple models using lift curves, statistical diagnostics and ROI metrics.
• Highly visual model comparison interface.
• Innovative Cutoff node examines to determine probability cutoff point(s) for binary targets.
• Report creation and distribution.
• Model result packages.
• Group processing for multiple targets and segments.
• Interactive environment for comparing and contrasting competing models and assessing the importance of key input variables on the predicted response times.
• Register Model node provides integrated environment for model registration into the SAS Metadata Server.
• Macro can also be used for registering models developed with SAS code into the SAS Metadata Server.

Automated scoring process

• Interactive scoring in a variety of real-time or batch environments.
• Automatically generates score code in SAS, C, Java and PMML.
• Score data based on models saved as PMML documents (experimental).
• Score SAS Enterprise Miner models directly inside Aster, EMC Pivotal (previously Greenplum), IBM DB2, IBM Netezza, Oracle and Teradata databases with SAS Scoring Accelerator.
• Model registration and management.
• Deploy models in multiple environments.
• Integrate SAS Enterprise Miner training and scoring processes directly into other SAS solutions.

Open, extensible design

• Extension node for easily adding tools and personalized SAS code.
• Interactive editor features for training and score code.
• Integrate text mining for analysis of both structured and unstructured data.
• Incorporate time series, Web paths and associations rules as additional input variables into the model development process.

Scalable processing

• The Java client and the SAS server architecture both scale from single-user to large enterprise solutions.
• Server-based processing and storage.
• Grid computing, in-database and in-memory processing options.
• Asynchronous model building.
• Ability to stop processing cleanly.
• Parallel processing.
• Multithreaded predictive algorithms.