Introduction to Machine Learning: Regression - Part 1

Dive into a comprehensive 51-minute webinar on regression analysis, the first part of an Introduction to Machine Learning series presented by Antoine Victoria and Wiktor Owczarz from Compass Lexecon's Data Science team. Explore the fundamentals of estimating relationships and predicting outcomes for continuous target variables. Learn essential regression concepts, algorithms, and techniques for measuring and evaluating variable relationships. Gain insights into practical considerations and potential pitfalls when applying regression, illustrated through various examples. Discover the basics of statistical learning, its applications, and the differences between parametric and non-parametric methods. Examine real-world applications, including simple and multiple regression models, with a focus on price vs. cost analysis and cartel overcharge estimation. Understand how to interpret regression results, assess model accuracy, and determine what constitutes a good model according to industry experts.

Intro
CONTENTS
ABOUT COMPASS LEXECON
WHAT IS STATISTICAL LEARNING?
VARIOUS APPLICATION OF STATISTICAL LEARNING
HOW-(NON)PARAMETRIC METHODS
WHEN DO WE CARE MORE ABOUT INFERENCE THAN PREDICTION
LINEAR REGRESSION AT COMPASS LEXECON
SIMPLE REGRESSION: PRICE VS COST
WHAT IS THE IMPACT OF COST ON PRICE?
SIMPLE LINEAR REGRESSION MODEL
ESTIMATION OF THE PARAMETERS BY LEAST SQUARES
MULTIPLE REGRESSION: CARTEL EXAMPLE
MULTIPLE REGRESSION AND CARTEL EXAMPLE-DUMMY VARIAB
CARTEL OVERCHARGE-BEFORE-DURING-AFTER COMPARISON
OVERCHARGE-CROSS-SECTIONAL COMPARISON
OVERCHARGE-DIFFERENCE-IN-DIFFERENCES ANALYSIS
INTERPRETING REGRESSION RESULTS
OVERALL MODEL ACCURACY-GOODNESS OF FIT
CONCLUSION: WHAT IS A GOOD MODEL (ACCORDING TO US)?