Course 805:
Introduction to Machine Learning

(3 days)


Course Description

With massive amounts of data (big data) available and inexpensive computing power to quickly process the data, it is now possible to find computational solutions to problems previously too expensive and time consuming to solve. This course starts by showing how to use machine learning (ML) as a support tool for solving business questions. It quickly moves on to provide an understanding of how ML works. Supervised models and their hyper-parameters are explored to include regression, ridge, lasso, svn, and neural networks (deep learning). Clustering of data is explored through unsupervised learning. Each section includes a complete working example showing how the ML model works. Exercises are included to allow students to explore how the model works.

Learning Objectives

  • Understand how supervised learning works
  • Know how to use linear regression models: regression, ridge, and lasso
  • Know how to use svn for non-linear problems
  • Know how to use classification models
  • Be able to use neural network models including deep learning for non-linear classification problems
  • Understand how to use unsupervised learning

Who Should Attend

Anyone wanting to understand supervised and unsupervised learning without spending weeks working through the math.

Course Outline


Overview of the Ecosystem of ML

  • Definitions of ML Terms
  • Taxonomy of ML

From Business Question to ML Question

  • Types of Questions ML Can Answer
  • Meta-Questions on the ML Question

Basic Supervised Learning Model

  • The Parts of an ML Model
  • Training the ML Model
  • Demonstration of Complete ML Project

Supervised Regression Models

  • Goodness of Fit Parameters
  • Outliers
  • Learning Rate
  • Multiple Features
  • Hands-On: Comparing Supervised Regression Models
  • Normalization and Standardization
  • Polynomial Regression
  • Overfitting
  • Regularization
  • Cross-Validation
  • Hands-On: Understanding Hyperparameters
  • Advanced Models: Ridge and Lasso

Supervised Classification Models

  • The Classification Model
  • Terms and Definitions
  • Binary Classification
  • Multiclass Classification
  • Solvers and Activation Function
  • Model Selection
  • Hands-On: Working with Classification Models

SVM Models

  • Why SVM?
  • Basic SVM
  • Kernel Tricks
  • Hands-On: Comparing Kernel Tricks

Decision Trees

  • Demonstration of a Decision Tree
  • Overview of Decision Tree Model
  • Hands-On: Decision Tree vs. Classification Regression
  • Combining Techniques: Ensembles and Forests
  • Hands-On: Working with a Forest

Supervised Neural Network

  • Limitations of Models Looked At
  • What Neural Network Provides
  • Working Through the Architecture
  • Hands-On: Working with the Architecture
  • Cost and Update Functions
  • Hands-On: Comparing Model and Hyperparameters
  • Demonstration of Advanced Models
  • Hands-On: Working with an Advanced Model

Unsupervised Learning: Clustering

  • What Is Different About Clustering?
  • KNN Clustering
  • Hands-On: Working with KNN Clustering
  • K-Mean Clustering
  • Hands-On: Working with K-Mean Clustering

Please Contact Your ROI Representative to Discuss Course Tailoring!