交通行動分析特論E

The role of models in urban transport planning Introduction to discrete choice models: •The Multinomial Logit model •Calculating effect sizes Parameter estimation methods Practical issues in discrete choice modeling (I): •Interaction effects •Aggregate forecasting techniques •Relevant statistical tests Reporting estimation results Practical issues in discrete choice modeling (II) •Aggregation and sampling of alternatives •Validation methods GEV models: •The Nested Logit model •The concept of consumer surplus •Other GEV models The Mixed Logit model •Mixed Logit formulation •Drawing from densities •Revisiting the concept of Value of Travel Time Machine learning and discrete choice •Introduction to classifiers x-AI (explainable AI)

Rather than one-way lectures, this course emphasizes hands-on learning, so the theory introduced in lectures will be applied via practice sessions and problem sets.

Attendance and class participation: 10% Practice session problem sets: 36% Final assignment : 54%

In principle, there are no pre-requisites for this class. However, a basic understanding of university level algebra, some calculus, and statistics (e.g. Wooldridge’s Introductory Econometrics) might facilitate things along the way.　 Experience in programming is not mandatory but it is desirable. We will provide a basic introduction to programming in R and some technical support for the code we share in class, but this is not a programming course, so keeping up with the course assignments depends on your effort. 講義の言語は英語ですが、日本語版の講義資料の提供及び日本語での学生対応は可能です。