SAS Institute Of Management Studies - Machine Learning Using SAS Viya 

Welcome to the Course!

Gettingstarted with Machine Learning using SAS® Viya® - In this module, you learn how you can meet today's business challenges with machine learning using SAS® Viya®. You start working on the project that runs throughout the course.

Introduction

Machine Learning in SAS Viya

Analytics Life Cycle

Case Study: Customer Churn

SAS Viya Tools for SAS Visual Data Mining and Machine Learning

Demo: Creating a Project

Predictive Modeling

Importance of Data Preparation

Essential Data Tasks

Dividing the Data

Addressing Rare Events Using Event-Based Sampling

Demo: Modifying the Data Partition

Managing Missing Values

Demo: Building a Pipeline from a Basic Template

SAS Viya in the SAS Platform: Architecture

WEEK 2 - Data Preparation and Algorithm Selection - In this module, you learn to explore the data and finish preparing the data for analysis. You also learn some general considerations for selecting an algorithm.

Introduction to Data Preparation and Algorithm Selection

Exploring the Data

Demo: Exploring the Data

Replacing Incorrect Values

Demo: Replacing Incorrect Values Starting on the Data Tab

Feature Creation

Text Mining

Demo: Adding Text Mining Features

Using Transformations to Handle Extreme or Unusual Values

Demo: Transforming Inputs

Selecting Useful Inputs

Demo: Selecting Features

Demo: Saving a Pipeline to the Exchange

Essential Discovery Tasks and Selecting an Algorithm

WEEK 3 - Decision Trees and Ensembles of Trees - In this module, you learn to build decision tree models as well as models based on ensembles, or combinations, of decision trees.

Introduction to Decision Trees and Ensembles of Trees

Basics of Decision Trees

Demo: Building a Decision Tree Model Using the Default Settings

Decision Trees for Categorical Targets: Classification Trees

Decision Trees for Interval Targets: Regression Trees

Improving the Decision Tree Model

Demo: Modifying the Structure Parameters

Recursive Partitioning

Splitting Criteria

Split Search

Demo: Modifying the Recursive Partitioning Parameters

Optimizing the Complexity of a Decision Tree Model

Pruning

Demo: Modifying the Pruning Parameters

Regularizing and Tuning the Hyperparameters of a Machine Learning Model

Building Ensemble Models

Perturb and Combine Methods

Bagging

Boosting

Comparison of Tree-Based Models

Demo: Building a Gradient Boosting Model

Forest Models

Demo: Building a Forest Model

WEEK 3 - Neural Networks - In this module, you learn to build neural network models. Introduction to Neural Networks

Beyond Traditional Regression: Neural Networks

Limitations of Neural Networks

Basics of Neural Networks

Estimating Weights and Making Predictions

Learning Process

Essential Discovery Tasks for Neural Networks

Demo: Building a Neural Network Using the Default Settings

Improving the Neural Network Model

Neural Network Architectures

Activation Functions

Shaping the Sigmoid

Demo: Modifying the Neural Network Architecture

Optimizing the Complexity of a Neural Network Model

Weight Decay

Early Stopping

Regularizing and Tuning the Hyperparameters of a Neural Network Model

Demo: Modifying the Learning and Optimization Parameters

WEEK 5 - Support Vector Machines - In this module, you learn to build support vector machine models.

Introduction to Support Vector Machines

Support Vector Machines as Classifier Models

Mathematical Definition of a Support Vector Machine

Maximum-Margin Hyperplane and Support Vectors

Essential Discovery Tasks for Support Vector Machines

Demo: Building a Support Vector Machine Using the Default Settings

Improving the Support Vector Machine Model

Optimization Problem

Accounting for Errors with Nonlinearly Separable Data

Demo: Modifying the Methods of Solution Parameters

Optimizing the Complexity of the Support Vector Machine Model

Feature Space Approach for Nonlinearly Separable Data

Kernel Trick

Demo: Increasing the Flexibility of the Support Vector Machine

Model Interpretability

Demo: Adding Model Interpretability

Regularizing and Tuning the Hyperparameters of the Support Vector Machine Model

WEEK 6 - Model Deployment - In this module, you learn how to select the model that best meets the requirements of your business challenge and put the model into production. You also learn about managing the model over time.

Introduction to Model Deployment

Essential Deployment Tasks

Selecting a Model

Numeric Measures of Model Performance

Confusion Matrix for Decision Predictions

ROC Charts and the C-Statistics

Charts Based on Response Rate: CPH and Lift

Ways of Comparing Models in Model Studio

Demo: Comparing Models within a Pipeline

Demo: Comparing Models across Pipelines

Demo: Exploring the Settings for Model Comparison and Selection

Scoring and Managing the Champion Model

Demo: Exploring the Features for Scoring and Managing a Model in Model Manager

Monitoring and Updating the Model