Predictive Analytics: Basic Modeling Techniques

About

What is Predictive Analytics? These methods lie behind the most transformative technologies of the last decade, that go under the more general name Artificial Intelligence or AI. In this course, the focus is on the skills that will allow you to fit a model to data, and measure how well it performs.

These skills also go under the names "machine learning" and "data science," the latter being a broader term than machine learning or predictive analytics but narrower than AI. This course is part of the Machine Learning Operations (MLOps) Program. We will be doing enough data science so that you get hands-on familiarity with understanding a dataset, fitting a model to it, and generating predictions. As you get further into the program, you will learn how to fit that model into a machine learning pipeline.

You will get hands-on experience with the top techniques in supervised learning: linear and logistic regression modeling, decision trees, neural networks, ensembles, and much more.

But most importantly, by the end of this course, you will know

• What a predictive model can (and cannot) do, and how its data is structured
• How to predict a numerical output, or a class (category)
• How to measure the out-of-sample (future)performance of a model

What you will learn

After completing this course, you will be able to:

• Develop a variety of machine learning algorithms for both classification and regression, including linear and logistic regression, decisions trees and neural networks

• Evaluate machine learning model performance with appropriate metrics

• Combine multiple models into ensembles to improve performance

• Explain the special contribution that deep learning has made to machine learning task

Skills you learn

• Forecasting
• Artificial Intelligence
• Predictive Analytics
• Supervised Learning
• Operations
• Artificial Neural Networks
• Logistic Regression
• Decision Tree Learning
• Data Science
• Machine Learning

Syllabus

Week 1 – Data Structures; Linear and Logistic Regression

• Classification and Regression
• Rectangular Data
• Regression
• Partitioning and Overfitting
• Illustration - Linear Regression (for verified users)
• Knowledge Check 1.1
• Logistic Regression
• Illustration - Logistic Regression (for verified users)
• Understand and Prepare Data
• Visualization
• CRISP-DM framework
• P-Values
• Knowledge Check 1.2
• Discussion Prompt #1 (for verified students, graded)
• Quiz #1 (for verified students, graded)
• Exercise #1 - Linear Regression (for verified students, graded)
• Exercise #2 - Logistic Regression (for verified students, graded)
• Summary

Week 2 - Assessing Models; Decision Trees

• Assessing Model Performance: Metrics
• ROC Curve and Gains Chart
• Decision Trees
• Illustration - Classification Tree (for verified users)
• Knowledge Check 2
• Quiz #2 (for verified students, graded)
• Exercise #3 - Regression Tree (for verified students, graded)
• Exercise #4 - Classification Tree (for verified students, graded)
• Summary

Week 3 – Ensembles

• Cross validation
• Module 3 Reading
• Ensembles
• Illustration - Ensemble Methods (for verified users)
• Knowledge Check 3
• Discussion Prompt #2 (for verified students, graded)
• Quiz #3 (for verified students, graded)
• Exercise #5 - Ensemble Methods (for verified students, graded)
• Summary

Week 4 - Neural Networks

• Neural Nets
• Illustration - Neural Nets (for verified users)
• Deep Learning
• Reading
• Knowledge Check 4
• Quiz #4 (for verified students, graded)
• Exercise #6 - Neural Nets (for verified students, graded)
• Summary

Auto Summary

Discover the essentials of Predictive Analytics with this 26-hour course by edX, focusing on data fitting and model performance. Ideal for those in Data Science & AI, it offers hands-on experience in understanding datasets and generating predictions. Suitable for beginners, with Starter and Professional subscription options available.