FOCUS AREAS

SAS Global Forum 2019 Super Demos

Super Demo Station 3

Day Time Title Abstract and Presenter
Sunday 4/28 4:30 PM Modeling Nonproportional Hazards by Performing Restricted Mean Survival Time Analysis Learn how to use the new RMSTREG procedure to do survival analysis based on the restricted mean survival time. (Changbin Guo)
Sunday 4/28 5:30 PM Longitudinal Data Analysis with the GEE Procedure Learn about analyzing longitudinal data with the GEE procedure and weighted GEE analyses for handling missing data due to dropouts. (Michael Lamm)
Monday 4/29 10:30 AM Modeling Data with Endogeneity Using SAS Econometrics Regression models based on observational data often suffer from endogeneity because of omitted confounder variables, measurement errors, simultaneity, or self-selection bias. Unless appropriate econometric techniques are used, endogeneity causes biased parameter estimates and incorrect inference. In this demo, you will learn how to use SAS Econometrics to correct for endogeneity and obtain unbiased estimates and standard errors. (Gunce Walton)
Monday 4/29 11:00 AM Addressing Excess Zeros Learn how to handle excess zeros in mixture model count data by using PROC FMM. (Dave Kessler)
Monday 4/29 11:30 AM Partial Proportional Odds Modeling with the LOGISTIC Procedure Learn how to use the LOGISTIC procedure to fit data by combining aspects of logistic. (Bob Derr)
Monday 4/29 12:00 PM New Optimization Features in SAS/OR 15.1 Learn about the optimization enhancements in SAS/OR 15.1, including improved solver performance, stability, and memory use; new nonlinear and network optimization solver algorithms; and PROC OPTMODEL access to the local search optimization solver. (Ed Hughes)
Monday 4/29 1:00 PM Estimating Causal Effects from Observational Data with PROC PSMATCH Learn how to use the PSMATCH procedure to do propensity score analysis and causal inference. (Michael Lamm)
Monday 4/29 1:30 PM Econometric Modeling with SAS Econometrics: Latest Modeling Trends and Package Enhancements This demo addresses the latest trends in econometric modeling and demonstrates their implementation in SAS Econometrics by using several examples from time series, cross-sectional, panel, and spatial data modeling. (Jan Chvosta)
Monday 4/29 2:00 PM Highlights of Model-Based Clustering Learn about model-based clustering, its advantages and disadvantages, and how to use PROC MBC to handle data of this nature. (Dave Kessler)
Monday 4/29 2:30 PM Parallel Computations for SAS/IML Programmers A new SAS Cloud Analytic Services (CAS) action, called the iml action, can execute programs that are written in the SAS/IML language. Learn how to use the action to perform certain computations in parallel on multiple threads. In conjunction with the ASTORE procedure, you can also deploy and score custom models that are written in the SAS/IML language. (Rick Wicklin)
Tuesday 4/30 10:00 AM PROC LOGSELECT and PROC GENSELECT: Old Friends, New Clothes Join us for an introduction to the LOGSELECT and GENSELECT procedures, based on SAS Viya, for logistic regression and generalized linear models, including highlights of each procedure’s features and tips and tricks for their use. (Dave Kessler and Bob Derr)
Tuesday 4/30 10:30 AM Assessing Predictive Accuracy of Survival Models with the PHREG Procedure Learn how to use the PHREG procedure to assess predictive accuracy of survival models using concordance statistics and time-dependent ROC curves. (Changbin Guo)
Tuesday 4/30 11:00 AM SAS Simulation Studio 15.1: New Features Learn about the new discrete-event simulation features in SAS Simulation Studio 15.1, including an extensively revamped user interface, much faster model execution, and easier authentication for batch mode runs of simulation models. (Ed Hughes)
Tuesday 4/30 11:30 AM Estimating Causal Effects from Observational Data with PROC CAUSALTRT Learn how to use the CAUSALTRT procedure to estimate treatment effects from observational data. (Michael Lamm)
Tuesday 4/30 12:00 PM Introducing the CAUSALGRAPH Procedure for Graphical Causal Model Analysis Learn how to use the CAUSALGRAPH procedure to find valid estimators for causal effects, even when only observational data are available. (Clay Thompson)
Tuesday 4/30 1:00 PM Modern Time Series Tools for Scenario Analysis and Portfolio Learn how to use two of the many modern time series tools in SAS Econometrics. Part 1 of this demo illustrates how to perform scenario analysis, often required in bank stress testing (CCAR), by using the Bayesian conditional forecast in the VARMAX procedure. Part 2 shows how to use the machine learning algorithm in the HMM procedure to recognize market states for portfolio optimization and determine a trading strategy that beats the market. (Xilong Chen)
Tuesday 4/30 1:30 PM Introducing the BGLIMM Procedure for Bayesian Generalized Linear Mixed Models Learn about the BGLIMM procedure, a new, high-performance, sampling-based procedure for Bayesian generalized linear mixed modes. (Amy Shi)
Tuesday 4/30 2:00 PM Accelerating Machine Learning through Simulation: A Case Study Come and see how the SAS Center of Excellence uses SAS Simulation Studio and machine learning together to deliver fast performance prediction and sensitivity analysis for clinical trial enrollment planning. (Bahar Biller)
Tuesday 4/30 2:30 PM What’s New in SAS/STAT 15.1 Hear about the highlights of this important new release. (Fang Chen)
Tuesday 4/30 3:00 PM The CAUSALMED Procedure for Causal Mediation Analysis Learn how to do causal mediation analysis with the CAUSALMED procedure. (Clay Thompson)
Tuesday 4/30 3:30 PM Easily Perform Competing-Risks Survival Analysis with SAS Studio Tasks Learn how to use SAS Studio tasks to quickly develop code for analyzing time-to-event data with competing risks. (Brian Gaines)

Super Demo Station 4

Day Time Title Abstract and Presenter
Sunday 4/28 4:30 PM Put a Label on It: Enrich Your Data with Semisupervised Learning In many applications, there is an abundance of data in the form of observations with attributes but no label assigned to say what those observations represent, a requirement for building a predictive model. Labeling observations takes a lot of time and domain expertise—it’s expensive. Come hear about a new capability in SAS Visual Data Mining and Machine Learning to provide labels for data by using semisupervised learning techniques. (James Cox)
Sunday 4/28 5:30 PM Introducing Pattern Matching for Graph Queries in SAS Viya 3.4 Want to execute graph queries that search for copies of a query graph within a larger graph? SAS Visual Data Mining and Machine Learning 8.3 on SAS Viya 3.4 provides this capability with the new patternMatch action and the PATTERNMATCH statement in the NETWORK procedure. Come see examples including social network search and fraud detection in transactional networks. (Matthew Galati)
Monday 4/29 10:30 AM Using Open Source Programming in Machine Learning Pipelines SAS, Python, and R are used by many data scientists, who often bring together teams with varied backgrounds and experiences to solve complex problems in new ways. This presentation focuses on how SAS is bringing these often-siloed worlds together in SAS Visual Data Mining and Machine Learning. (Jesse Luebbert)
Monday 4/29 11:00 AM Executing SAS Code in SAS Visual Data Mining and Machine Learning Pipelines Whether your goal is custom data preprocessing, modeling, or data summarization and visualization, the SAS Code node gives you ultimate flexibility to customize your machine learning pipelines to suit your needs. This presentation walks you through a demonstration of different ways to use this node and highlights other examples available to you today in the SAS GitHub repository. (Wendy Czika)
Monday 4/29 11:30 AM Integrating SAS Visual Data Mining and Machine Learning Models into SAS Enterprise Miner Projects The SAS Viya node in SAS Enterprise Miner can be used to create models based on SAS Visual Data Mining and Machine Learning and integrate them into a SAS Enterprise Miner project. Come see how the SAS Viya node provides a template that gets you started with code needed to run procedures in the SAS Viya execution environment from SAS Enterprise Miner. (Jagruti Kanjia)
Monday 4/29 12:00 PM Scalable Cloud-Based Time Series Analysis and Forecasting The TSMODEL procedure, together with its various packages, provides all the functionalities of the TIMESERIES procedure and many additional cutting-edge data analysis tools and useful utilities. PROC TSMODEL greatly accelerates BY-group processing because the BY groups are processed concurrently on multiple nodes of the SAS Viya distributed environment. This demo provides an overview of the major features of PROC TSMODEL. (Michael Leonard)
Monday 4/29 1:00 PM What's New in SAS Visual Forecasting 8.4 This demo illustrates the following new or improved functionalities in SAS Visual Forecasting 8.4: derived attributes that are integrated with faceted search, attributes-based filters, new exploration and navigation capabilities, new modeling strategies, greatly improved segmentation, and more. (Jerzy Brzezicki)
Monday 4/29 1:30 PM Replace Your ARIMA Transfer Function Models with UCM Transfer Function Models! Starting with SAS/ETS 15.1, the UCM procedure enables you to specify transfer function relationships in an unobserved components model (UCM). You can now use PROC UCM to fit both the UCM and ARIMAX models, as well as many variations of these models. This demo uses real-world examples to illustrate this new feature. (Michael Leonard)
Monday 4/29 2:00 PM Scoring Predictive Models with SAS/STAT Software Learn about methods for scoring predictive models using SAS/STAT, including the CODE and SCORE statements and the PLM procedure. (Phil Gibbs)
Monday 4/29 2:30 PM Build Fully Connected Neural Networks with SAS Studio Tasks and SAS Viya Learn how to quickly develop code for training fully connected neural networks by using SAS Studio tasks with SAS Viya. (Brian Gaines)
Tuesday 4/30 10:00 AM A Programmatic Approach to Automated and Interpretable Machine Learning This presentation introduces a programmatic approach in SAS Visual Data Mining and Machine Learning that supports (a) automated feature engineering, (b) intelligent model selection, and (c) interpretation of model results. (Ilknur Kaynar Kabul)
Tuesday 4/30 10:30 AM A Visual Approach to Interpretability of Machine Learning Models In this demo, you will learn various ways to reliably explain black-box models created by SAS Visual Data Mining and Machine Learning pipelines. The new interpretability reports are enhanced by simpler explanations that are generated by natural language processing. Examples include the explanation of complex models that range from high-stakes decisions in health care to loan approvals. (Funda Gunes)
Tuesday 4/30 11:00 AM Deploying Models to Micro Analytics Service Learn how to deploy models you’ve built in Visual Data Mining and Machine Learning to the Micro Analytics Service in order to score streaming data using a RESTful interface. (Shawn Pecze)
Tuesday 4/30 11:30 AM It’s Opportune to Autotune Come see the latest advancements in automated model tuning, including searching for the best models across multiple model types all at once and tuning on the basis of multiple objectives. (Brett Wujek)
Tuesday 4/30 12:00 PM Independent Component Analysis Using the ICA Procedure Learn how to use the ICA procedure to perform independent component analysis that reveals the hidden structure of data. (Ning Kang)
Tuesday 4/30 1:00 PM Using Open Programming Interfaces in SAS Visual Forecasting Master the power of SAS Visual Forecasting from your favorite language! We will show you how to access the powerful distributed forecasting engine of Visual Forecasting from SAS, Python, and R. (Michele Trovero)
Tuesday 4/30 1:30 PM Override External Forecasts in SAS Visual Forecasting Even if you don't use Model Studio to create forecasts, you can use its advanced override functionality to adjust your final forecasts. (Jerzy Brzezicki)
Tuesday 4/30 2:00 PM Tips and Tricks for Information Extraction from Textual Data Information extraction is useful for transforming unstructured data into new, structured data variables. This demo gives you tips and tricks that make information extraction easy by using SAS Visual Text Analytics. Examples show you how to handle negation, identify n-grams, and split longer documents. (Teresa Jade)
Tuesday 4/30 2:30 PM What’s New in SAS Visual Text Analytics 8.4 SAS Visual Text Analytics provides a modern, flexible, and comprehensive analytics framework to harness the power of natural language processing (NLP), linguistic rules, machine learning, deep learning, and search. This demo gives you a sneak preview of the upcoming release of SAS Visual Text Analytics in action. (Jared Peterson)
Tuesday 4/30 3:00 PM Using DLPy to Import Open Source Deep Learning Models SAS provides an easy-to-use Python package called DLPy for automating common deep learning tasks. In this demo, you learn to import an open source model and use the resulting DLPy model for image classification. (Doug Cairns)
Tuesday 4/30 3:30 PM A Programmatic Approach to Text Analytics This presentation introduces a programmatic approach in SAS Visual Text Analytics. Learn to programmatically solve text analytics tasks like document summarization, topic extraction, and automatic rule generation by using SAS Studio. (Jim Cox)