Machine Learning Basics With Minitab

Theory with elaborated examples and Minitab tutorials

What you'll learn

You will learn the fundamentals of machine learning with a focus on practical applications using Minitab.

You will also learn how to apply these techniques to real world problems in a wide variety of application areas.

This hands-on approach will give you the confidence and skills you need to succeed in a career in data analysis or machine learning.

By the end of the course, you'll be able to build and implement regression and classification models and gain a deep understanding of their underlying concepts.

Requirements

Basic knowledge in Statistics.

It is recommended to use this version because earlier versions cannot read the attached Minitab project files. However, the tutorial and example data files can also be downloaded in Excel *.xlsx format, so that students with earlier Minitab versions can follow the course and do the exercises on their own.

No programming skills.

Description

Course Title: Machine Learning Basics with MinitabCourse Description:This comprehensive course is designed to provide a detailed understanding of the basics of machine learning using Minitab, with a focus on supervised learning. The course covers the fundamental concepts of regression analysis and binary logistic classification, including how to evaluate models and interpret results. The course also covers tree-based models for binary and multinomial classification.The course begins with an introduction to machine learning, where students will gain an understanding of what machine learning is, the different types of machine learning, and the difference between supervised and unsupervised learning. This is followed by an overview of the basics of supervised learning, including how to learn, the different types of regression, and the conditions that must be met to use regression models in machine learning versus classical statistics.The course then delves into regression analysis in detail, covering the different types of regression models and how to use Minitab to evaluate them. This includes a thorough explanation of statistically significant predictors, multicollinearity, and how to handle regression models that include categorical predictors, including additive and interaction effects. Students will also learn how to make predictions for new observations using confidence intervals and prediction intervals.Next, the course moves onto model building, where students will learn how to handle regression equations with "wrong" predictors and use stepwise regression to find optimal models in Minitab. This includes an overview of how to evaluate models and interpret results.The course then shifts to binary logistic regression, which is used for binary classification. Students will learn how to evaluate binary classification models, including good fit metrics such as the ROC curve and AUC. They will also use Minitab to analyze a heart failure dataset using binary logistic regression.The course then covers classification trees, including an overview of node splitting methods such as splitting by misclassification rate, Gini impurity, and entropy. Students will learn how to predict class for a node and evaluate the goodness of the model using misclassification costs, ROC curve, Gain chart, and Lift chart for both binary and multinomial classification.Finally, the course covers the concept and use of predefined prior probabilities and input misclassification costs, and how to build a tree using Minitab. Throughout the course, students will gain hands-on experience applying the concepts learned in real-world scenarios.Overall, this course provides a thorough understanding of machine learning basics using Minitab, with a focus on supervised learning, regression analysis, and classification. Upon completion of this course, students will have the knowledge and skills to apply supervised machine learning techniques to real-world data problems.

Overview

Section 1: Introduction

Lecture 1 Introduction to Supervised Machine Learning

Section 2: Regression in Classical Statistics and in Machine Learning (ML)

Lecture 2 Introduction to Regression

Lecture 3 Evaluating Regression Models

Lecture 4 Conditions for Using Regression Models in ML versus in Classical Statistics

Lecture 5 Statistically Significant Predictors

Lecture 6 Regression Models Including Categorical Predictors. Additive Effects

Lecture 7 Regression Models Including Categorical Predictors. Interaction Effects

Lecture 8 Multicollinearity among Predictors and its Consequences

Lecture 9 Prediction for New Observation. Confidence Interval and Prediction Interval

Lecture 10 Model Building. What if the Regression Equation Contains "Wrong" Predictors?

Lecture 11 Stepwise Regression and its Use for Finding the Optimal Model in Minitab

Section 3: Regression with Minitab. Examples and Exercises

Lecture 12 Regression with Minitab. Example. Auto-mpg. Part 1

Lecture 13 Regression with Minitab. Example. Auto-mpg. Part 2

Section 4: Regression Tree Models

Lecture 14 The Basic Idea of Regression Trees

Section 5: Regression Trees with Minitab. Examples and Exercises

Lecture 15 Regression Trees with Minitab. Example. Bike Sharing. Part 1

Lecture 16 Regression Trees with Minitab. Example. Bike Sharing. Part 2

Section 6: Classification by Binary Logistic Regression Models

Lecture 17 Introduction to Binary Logistic Regression

Lecture 18 Evaluating Binary Classification Models. Goodness of Fit Metrics. ROC Curve. AUC

Section 7: Binary Logistic Regression Models with Minitab. Examples and Exercises

Lecture 19 Binary Logistic Regression with Minitab. Example. Heart Failure. Part 1

Lecture 20 Binary Logistic Regression with Minitab. Example. Heart Failure. Part 2

Section 8: Classification Tree Models

Lecture 21 Introduction to Classification Trees

Lecture 22 Node Splitting Methods 1. Splitting by Misclassification Rate

Lecture 23 Node Splitting Methods 2. Splitting by Gini Impurity or Entropy

Lecture 24 Predicted Class for a Node

Lecture 25 The Goodness of the Model - 1. Model Misclassification Cost

Lecture 26 The Goodness of the Model - 2. ROC. Gain. Lift. Binary Classification

Lecture 27 The Goodness of the Model - 3. ROC. Gain. Lift. Multinomial Classification

Lecture 28 Predefined Prior Probabilities and Input Misclassification Costs

Lecture 29 Building the Tree

Section 9: Classification Tree Models. Examples and Exercises with Minitab

Lecture 30 Classification Trees with Minitab. Example. Maintenance of Machines. Part 1

Lecture 31 Classification Trees with Minitab. Example. Maintenance of Machines. Part 2

Section 10: Comprehensive Project 1. Regression Models for New York Yellow Taxi Trips

Lecture 32 Data Cleaning. Part 1

Lecture 33 Data Cleaning. Part 2

Lecture 34 Creating New Features

Lecture 35 Polynomial Regression Models for Quantitative Predictor Variables

Lecture 36 Interactions Regression Models for Quantitative Predictor Variables

Lecture 37 Qualitative and Quantitative Predictors. Interaction Models

Lecture 38 Final Models for Duration and TotalCharge. Without Validation

Lecture 39 Underfitting or Overfitting. The "Just Right" Model

Lecture 40 The "Just Right" Model for Duration

Lecture 41 The "Just Right" Model for Duration. A More Detailed Error Analysis

Lecture 42 The "Just Right" Model for TotalCharge

Lecture 43 The "Just Right" Model for TotalCharge. A More Detailed Error Analysis

Lecture 44 Regression Trees for Duration and TotalCharge

Section 11: Comprehensive Project 2. Classification Models for Predicting Learning Success

Lecture 45 Predicting Learning Success. The Problem Statement

Lecture 46 Predicting Learning Success. Binary Logistic Regression Models

Lecture 47 Predicting Learning Success. Classification Tree Models

Section 12: Open Access Machine Learning Repository for Self-study and Self-practice

Lecture 48 Open Access Machine Learning Repository for Self-study and Self-practice

This course is designed for students with a basic statistics background who are new to machine learning and want to gain practical skills in this field. No programming experience is necessary, but the course will introduce you to the advanced use of Minitab's menu-driven interface. Machine Learning is a multi-disciplinary field, often only to be learned in more depth over several books and courses, but this course is the perfect first learning resource.