In this course we will concentrate on the study and practice of info-metrics modeling and inference. We will concentrate on estimation and inference of problems where the information we have is quite limited and often very noisy. Though similar problems arise across most disciplines, we will focus on the study of Information-Theoretic (IT) methods of inference in general (within an interdisciplinary perspective) but with a strong emphasis on problems in the social sciences and economics.
We will emphasize both the fundamental theory, the motivation for using the theory, its background, and practice the theory with real or artificial data. Part of the lectures will be complemented with computer experiment in class. We will compare the info-metrics framework with other methods, like the maximum likelihood and least squares.
For further background a web support with many example, references, codes and software is available. See: http://info-metrics.org/
Target group and learning outcomes
The course is beneficial to graduate students, researchers and academics from across disciplines with an interest in solving all types of empirical problems with complicated data and with minimal imposed structure and statistical assumptions. Examples include modelling and inference of problems with small, ill-behaved or complex data.
After a successful completion of this course participants are expected to be able to:
- Understand the basic info-metrics framework, its motivation and background and under what conditions to use it;
- Construct and use info-metric models for solving applied policy and other problems
- Understand the differences between info-metric, maximum likelihood and least squares approaches;
- Apply info-metric estimation techniques to real world problems;
- Understand the class of Information-Theoretic (IT) methods of inference.
- Perform diagnostics and tests of info-metric (and other information-theoretic) methods.
Assumed prior knowledge
The background needed for the course is statistics and/or econometrics traditionally studied during the first year of graduate school in any quantitative discipline.
For PhDs of WASS there is a fee of 125 euros. For PhD students from other Wageningen graduate schools the fee is 250 euros. For all other participants and for staff members (fellows/post docs), there is a fee of 500 euros for the whole course (including drinks, lunches, and a course dinner).
Morning sessions are scheduled from 9.00-12.00 and will be mostly used to discuss theory. In most afternoons practical sessions are scheduled from 13.30-16.00 in which participants get hands-on experience in applying info-metrics techniques using statistical, econometric or other publicly available software. (Note, that depending on class material some computer experiments will be done during the morning sessions, and some more theoretical lectures in parts of the afternoon sessions). There will be a course dinner Thursday evening.
Outline of the Course in Hours
The entire course, including preparation, homework and a small final project consisting of completing a problem set with both theoretical and computational problems involves 3 ECTS (84 hours). Credits can only be obtained by completing the final project. People should start working on the problem set at the end of the course and it is due a eight weeks after the end of the sessions (to be submitted to Golan).
The course will be composed of lectures, open discussions, and complementing exercises (to be completed after the course).
The exercises and computer practice will allow each participant to gain the most out of this course where a substantial amount of computing and practice is necessary. Those who are used to write their own computer codes, the computing can be done by using different software, such as Matlab, GAMS, Python, R, etc. For those who wish to use common statistical/econometric software, the methods we discuss in this course can be used within some of the main software packages, such as STATA, SAS and NLOGIT (LIMDEP). The basic codes will be provided to the participants and are available on the main book’s web page.
Temporary licenses for GAMS and NLOGIT will be provided to the participants (STATA and SAS examples will be provided as well).
A brief tentative timeline and topics
| May 6 (9:00-12:00) Introduction, Rational Inference and Metrics of Info-Metrics
|| (Chapters 1 – 3)
| May 6 (13:30-16:00) Maximum Entropy (Part I).
|| (Chapter 4)
| May 7 (9:00-12:00). Maximum Entropy (Part II); Examples; Test Statistics and Diagnostics
|| (Chapters 4 – 6; Note Chapters 5 – 6 are examples)
| May 7 (13:30-16:00) Lab experiments and practice
| May 8 (9:00-12:00) Prior Information; Possibly some computer experiments.
|| (Chapter 8)
| May 8 (13:30-16:00) Computer Lab Experiments; The Info-Metrics Framework (Part I)
|| (Chapter 9)
| May 9 (9:00-12:00) The Info-Metrics Framework and Examples (Part II)
|| (Chapter 9)
| May 9 (13:30-16:00) Information Theoretic Methods (Part I – Discrete Choice); Lab Exercises
|| (Chapter 12)
| May 10 (9:00-12:00) Information Theoretic Methods (Part II – Continuous Problems); Lab Exercise
|| (Chapter 13)
| May 10 (13:30-16:00); Computer Lab; Empirical Examples; Class Summary
|WASS, PE&RC and WIMEK/SENSE PhDs with TSP
|a) All other PhD candidates b) Postdocs and staff of the above mentioned Graduate Schools
Fee includes coffee/tea, lunch and course materials.
NB: for some courses, PhD candidates from other WUR graduate schools with a TSP are also entitled to a reduced fee. Please consult your Education/PhD Programme Coordinator for more information
The participants can cancel their registration free of charge 1 month before the course starts. A cancellation fee of 100% applies if a participant cancels his/her registration less than 1 month prior to the start of the course.
The organisers have the right to cancel the course no later than one month before the planned course start date in the case that the number of registrations does not reach the minimum.
The participants will be notified of any changes at their e-mail addresses.