Discrete choice, part 3: Location and scale normalization in logit models

preview_player
Показать описание
I discuss identification in the conditional and multinomial logit models respectively. In particular, I prove why we cannot estimate an intercept in the conditional logit model and why we must normalize the betas for one of the alternatives in the multinomial logit model. Finally, I show why we must normalize the variance in the multinomial logit model (the proof is the same for conditional logit).

In more general terms, these types of identification normalizations are referred to as "scale" and "location" normalization in discrete choice models.

Chapters
00:00 Intro
00:23 Location normalization in conditional logit
3:32 Location normalization in multinomial logit
7:20 Scale normalization in multinomial logit
  1. Anders Munk-Nielsen
  2. Econometrics
  3. Logit
  4. Normalization
  5. Identification
  6. Conditional logit
Рекомендации по теме
Discrete choice, part 0:10:20

Discrete choice, part 3: Location and scale normalization in logit models

Part 1/3: Measuring 0:05:13

Part 1/3: Measuring willingness-to-pay using discrete choice experiments

Common mistakes in 0:06:41

Common mistakes in discrete choice modeling. Episode 3: Aggregate elasticities

Beyond Discrete Choice 0:54:06

Beyond Discrete Choice in Health: Multiple Discrete-Continuous Extreme Value (MDCEV) Models

Part 3/3: Mechanism 0:07:01

Part 3/3: Mechanism of new method separated adaptive dual response

3-3-Discrete models of 0:34:11

3-3-Discrete models of sample data part 3

Andrew Tomkins: Models 0:31:28

Andrew Tomkins: Models for Discrete Choice

Discrete Choice Analysis: 0:03:56

Discrete Choice Analysis: Causal Inference Bootcamp

Discrete choice, part 0:13:06

Discrete choice, part 1: Conditional logit and multinomial logit

Confounders in Discrete 0:03:03

Confounders in Discrete Choice Analysis: Causal Inference Bootcamp

Lecture 9 2:46:32

Lecture 9 Discrete Choice Model 1

Week 3: Random 0:11:39

Week 3: Random Utility Model | Video 1: Discrete Choice

Discrete choice analysis 0:00:34

Discrete choice analysis Meaning

DISCRETE CHOICE MODEL 0:09:46

DISCRETE CHOICE MODEL

13 Binary discrete 0:00:43

13 Binary discrete Choice model

What are... Discrete 0:13:01

What are... Discrete Choice Experiments? Matthew Quaife

Performance of a 0:01:25

Performance of a comprehension question in discrete-choice experiment surveys (DCE)

Introduction to Discrete 0:03:03

Introduction to Discrete Choice Models

Using the Choice 0:03:11

Using the Choice Model

Maximum Likelihood estimation 0:04:00

Maximum Likelihood estimation - an introduction part 3

Common mistakes in 0:16:51

Common mistakes in discrete choice modeling. Episode I: To be or not to be significant

Part 2/3: Shortcomings 0:06:23

Part 2/3: Shortcomings with the estimation of willingness to pay

Nlogit - Import 0:02:33

Nlogit - Import data (.csv) and run Discrete choice modelling

Part 1 - 0:02:57

Part 1 - Discrete Choice Analysis: Predicting Individual Behavior and Market Demand

Комментарии
Автор

Great explanation! Thank you very much!

tallyskalynkafeldens
Автор

Nice video!! Best source for this proof, and I checked several books. Thank you!

lucasng
Автор

Hello, professor! In the last part of the video, "MNL with dispersion", shouldn't the right side of the inequality be "beta_h * x_i", not "beta_h * x_h"? Thanks for the video, by the way. It's really helpful!

yunsunpark