You are planning to attend AnalyticsX next month, aren't you? If not, you're missing out on some amazing presentations!
Below you will find a list of SAS presenters and their topics. To get the latest details, including times and room locations, visit the session catalog.
Collaborative, Open and Automated Machine Learning using SAS Visual Data Mining and Machine Learning
Presenter: Jonathan Wexler
The application of machine learning is pervasive. Whether you are preparing data for visualization, exploring data for hidden insights, or building model pipelines - you should be able to use your experience of choice. In this session you will learn how SAS Visual Data Mining and Machine Learning enables automated insights, interactive/collaborative machine learning, open source integration and sas programming - all within one platform. Not a Data Scientist -- no problem! You will learn how to take advantage of best practices from other users, or even become the expert and share insights and best practices with others. Learn about this solution and see it in action.
Making analytics pervasive in an organization
Presenters: Saurabh Gupta, Udo Sglavo
From the world leader in analytics comes a new release that adds newer capabilities to help make analytics pervasive in an organization. It provides a single, open, cloud-ready environment that serves everyone from data scientists to business analysts, application developers to executives. Explore relationships and patterns in data smartly, quickly and easily. Illuminate critical insights with self-service analytics. Analyze different types of data, with unseen granularity. Boost your analytical productivity through collaboration. Automate modeling and deployment. Learn how you can put analytics into action.
SAS® Visual Forecasting - Version 8.3: An Overview of New Forecasting Features
Presenter: Joe Katz
SAS Visual Forecasting is the next generation of SAS product for forecasting based on the SAS(r) Viya(tm) architecture. It features a new graphical interface centered on the use of pipelines, a new midtier that uses microservices, and a new server environment based on the Cloud Analytics Services (CAS). It provides end-to-end capability to explore and prepare data, apply various modeling strategies, compare forecasts, override statistical forecasts, and visualize results. The process workflow to model generation and forecasting is centered on the use of pipelines and is shared with SAS(r) Visual Data Mining and Machine Learning and SAS(r) Visual Text Analytics. Collaborate within your team or across your organization by sharing reusable components - nodes, pipeline templates, and projects. New to this release are functionality for segmentation, calendar events, importing events, modifying event usage in modeling nodes, and three modeling strategies incorporating Neural Networks. In addition, Forecast Analysts/Data Scientists can access the power of SAS Visual Forecasting though a flexible and powerful programming environment. Please join us in this session for a demonstration of this latest release.
Playing Favorites: My Top 10 Features in SAS® Visual Data Mining and Machine Learning’s Model Studio
Presenter: Wendy Czika
In SAS® Visual Data Mining and Machine Learning, Model Studio provides a pipeline-centric, collaborative modeling environment. This environment enables you to chain together steps for preprocessing your data by modifying the data or performing feature engineering, using supervised learning algorithms to make predictions, and then using the assessment measure of your choice to compare models. Once the champion model has been determined, you can deploy the model that the pipeline represents to score new data. Even with the second release of Model Studio coming right on the heels of the first release, many enhancements were made. These enhancements include nodes for running open source code, batch code from SAS® Enterprise Miner™, or score code that is generated outside Model Studio. Integration with other visual environments has also been enhanced to allow for seamless exploration and visualization of your data as you build models. Although it is hard to limit my favorites, I will present my top 10 list of the latest and greatest Model Studio features available in SAS Visual Data Mining and Machine Learning 8.3 for solving your biggest problems and increasing your productivity.
Neural-Network-Based Forecasting Strategies in SAS® Viya®
Presenter: Steven Mills
SAS® Viya® 3.4 includes new forecasting strategies that use neural networks (NNs) for modeling. These modeling approaches provide alternate strategies for large data and for capturing nonlinear interactions that might not be accurately portrayed by time series models.
The Panel Series Neural Network node models the response variable using a feedforward neural network and generates input features, such as lags, seasonal dummy variables, exponential smoothing models of the dependent variable, and trend components. The Stacked Modeling node first generates a panel series neural network model, and subsequently fits a time series model to the residuals. The resulting models are then combined to leverage the advantages of both the NN approach and the time series approach. Finally, with the Multistage Modeling node, you can use regression or an NN for modeling the lower level of a hierarchy, and use time series models for the upper levels. Then you can reconcile the forecasts to achieve the final forecast.
Have Your Cake and Eat It Too—With Python, R, and SAS®
Presenter: Jesse Luebbert
SAS, Python, and R software are used by many data scientists, who often bring together teams with varied backgrounds and experiences to solve complex problems in new ways. However, the mix of languages can lead to challenges within an enterprise analytics deployment, especially challenges of scalability, collaboration, and governance. This talk will focus on how SAS is alleviating those challenges and bringing these often-siloed worlds together in SAS(r) Viya(r). Concrete examples show where it is advantageous to bring SAS into R and Python clients or to bring R and Python into SAS clients and show how you can use Model Studio in SAS(r) Visual Data Mining and Machine Learning to make the best of both worlds.
Deploying SAS Viya Machine Learning Models Using SWAT, Python and Flask
Presenter: Jesse Luebbert
After machine learning models are built, often it can take a long time to get models into production, whether the model is meant to run as a batch job or score new records on a case by case basis. SAS gives you the ability to do both quickly and easily through score code and APIs. For scoring via APIs, Python’s Flask micro framework has become a popular way of getting applications up and running quickly, either for production or internal use. This talk will introduce Flask as a method of deployment and show how you can easily take machine learning models built in any SAS Viya interface and deploy them in Flask through SWAT, our Python interface to the SAS Cloud Analytic Services (CAS) engine.
