This course combines theory and practice to immerse you in the core concepts of neural network models and the essential practices of real-world application. During the course, you programmatically build a neural network and discover how to adjust the model’s essential parameters to solve different types of business challenges. You implement early stopping, build autoencoders for a predictive model, and perform an intelligent automatic search of the model hyperparameter values. The last lesson introduces deep learning. You gain hands-on practice building neural networks in SAS 9.4 and the cutting-edge cloud computing platform for big data analytics, SAS Viya.

Lernen Sie, wie Sie / Learn how to

• Programmatically build neural networks in SAS 9.4 and SAS Viya.
• Modify neural networks' parameters for better performance.
• Conduct automatic search for neural networks' hyperparameters through genetic algorithm.
• Enhance data with autoencoders and synthetic observations.

Zielgruppe / Who should attend

Those interested in learning about neural networks, general machine learning and data science techniques, and SAS software

Verfügbare Schulungsformen / Formats available	Dauer
Präsenzkurs:	2.0 Tage
Live Web Class:	14 Stunden

Before taking this course, you should have the following:

• Some familiarity with programming in SAS or SQL (or both).
• An understanding of predictive modeling.
• A basic understanding of calculus.

In diesem Kurs wird mit folgenden Software Modulen gearbeitet: SAS Viya Software

Neural Networks: Essentials

• Introduction.
• Multilayer perceptrons.
• Neural network modeling paradigm.
• Using a surrogate model to interpret neural network predictions.
• Other considerations.

Neural Network Details

• Parameter estimation.
• Numerical optimization methods.
• Regularization.
• Unbalanced data.
• SAS search optimizations (self-study).

Tuning a Neural Network

• Selecting hyperparameters with autotuning.

Introduction to Deep Learning

• Introduction to deep learning.
• Autoencoders.

Radial Basis Function Networks (Self-Study)