, Methods in Transportation Econometrics and Statistics (Master)
TU Dresden » ... » Professur für Ökonometrie und Statistik, insbes. im Verkehrswesen » Studium » VL Verkehrsökonometrie
PROFESSUR FÜR ÖKONOMETRIE UND STATISTIK, INSB. IM VERKEHRSWESEN

Home
Dr. Treiber

Tutorials

Lecture
slides

Vorlesungs-
skript

Example
surveys

Further
Reading

Statistik-
Formel-
sammlung

old exams/alte Klausuren

Interactive
Traffic
Simulation

Methods in Transportation Econometrics and Statistics (Master), Winter 2020/21

News

(2021-02-14) Solutions to the exam are available

(2021-01-20) Last Tutorial 11: Input-output model uploaded

(2021-01-20) Tutorial videos for the Lectures 5 and 6 and the video of the last Lecture 11 uploaded

(2021-01-15) Last lecture uploaded as pdf. The video will be uploaded next Tuesday

(2021-01-13) Information on the exam is now available

(2021-01-10) Lecture 10: Advanced Concepts of Discrete-Choice Theory as pdf and video

(2021-01-05) Lecture 09: Inferential Statistics of Discrete-Choice Models available as pdf and video

(2020-12-19) Lectures 08: Logit and Probit Models available as pdf and as video

(2020-12-15) Lectures 07: Basics of Discrete-Choice Models" available as pdf and as video

(2020-12-15) Lectures 06: "Obtaining the Data" as pdf and as video

(2020-12-01)
(2020-11-23) New tutorials 3 and 4 as pdf and as video

(2020-11-23) Lecture 05: Bayesian Approach to Inferential Statistics as pdf and video

(2020-11-17) 4th lecture online as video and pdf

(2020-11-11) Third lecture (second regular one) online as video and pdf

(2020-11-10) Solutions for Tutorials 01a and 02 available as pdf and video

(2020-11-06) Tutorial for the second lecture (Corona) is online

(2020-11-05) Video of the second lecture (Corona) is online

(2020-11-05) Second lecture online as pdf

(2020-11-02) Solutions for Tutorial 01 available as pdf and video. For the video, see my Youtube channel

(2020-10-29) Youtube channel virtual lectures set up. First lecture (direkt link) is online

(2020-10-27) New detailled set of slides for the first lecture. The lecture video itself will be uploaded tomorrow

(2020-10-23) Until further notice, the tutorial will be held online (asynchronous) as well due to the present surge in Covid-19 cases.

The problems of the next couple of weeks are already online. At the time of the tutorial (every Friday, starting Oct 30, 2020), I will upload some video on the problem (in English), and also pdf's of the solutions to it (in English and German).

Examination

Due to the known Covid-xx circumstances, this semester's examination is in electronic form. In order to access the examination, please make sure you have registered at the TUD's examination platform. Links to the actual exam group will be given later.

Times

  • January 26, 10:00 trial exam (Probeklausur)
  • February 09, 10:00 real exam
The trial exam is there to make sure that everybody can log in and understand how the novel exam form works. Your results will not count for the real exam but surely it helps to pass the real exam more easily

Which kind of exam problems will you encounter?

Because of the limitations of this form of examination, the exam problems will be shorter and less heavy on math than in the past written exams all problems are given in English, only

They are more like aural examination problems and come in four technical forms
  • single or multiple choice (click one or more items in an offered list)
  • small numerical calculation problems where you enter a number in a textbox (a precision of three significant digits is enough)
  • analytical problems where you enter a simple formula such as sqrt(a^2+b^2)
  • text problems where you type in free text in a bigger box

Typical examples

Some of the examples for an aural examination for the Bachelor's course (German) are useful. Other examples follow:
  • Describe, in one sentence, the meaning of calibration and validation
  • In the last years, the total number of sold electric vehicles (EVs) could be well described by a model of unlimited growth.
    1. Is this statement well delimited in a statistical sense? Think of the spatial, temporal, and attribute delimitation of the population of electric vehicles
      [Text answer: spatial: no: unclear whether, e.g., Germany or the world temporal: OK, attributes: It is not mentioned whether only battery-electric vehicles count or whether plugin hybrids are included as well]
    2. In 2018, the model predicts 12 000 sold vehicles, in 2019 20 000. Give the number predicted in 2020 [enter the number in a box]
    3. If the number of sold electrical vehicles per year grows exponentially, is this also true for the sold total number since the dawn of time?
      • yes
      • no
  • For a binary-choice situation, the utilities gor the two alternatives have been calculated as V1=-2 and V2=-1 for the Logit model
    1. Tick all that applies: Without changing the choice probabilities, you can ...
      • multiply both V1 and V2 by a common positive factor
      • set V1=-1 and V2=0
      • multiply the non-symmetric gumbel-distributed random utilities ε1 and ε2 by -1
      • multiply V1, V2, and both εi by a common positive factor
      • exponentiate V1, V2, and both εi and replace the additive noise by a multiplicative one
    2. Calculate the probability for choosing alternative 1 [enter a number in a box]
    3. For a binary i.i.d. Probit model, the deterministic utilities for the calibrated model are V1=-1.6 and V2=-0.8
      • Why is the absolute value of the calibrated model expected to be smaller? [enter text]
      • Calculate the choice probability for alternative 1 [enter a number in a box]

Dates and Times


In this semester, the lecture is online (asynchronous) in form of videos. Further material for reading and links can be found below.

Until further notice, the tutorial will be held online (asynchronous) as well due to the present surge in Covid-19 cases



Youtube channel of this course

Materials for Download

Curriculum

Lecture Slides (English)

Vorlesungsfolien (Deutsch)

Vorlesungsskript (in German)

Notice: The English (not the German) slides will be complete enough so that the German lecture notes below will not be needed for the English speaking students

Das Skript ist teils ausführlicher als der in der Vorlesung behandelte Stoff. Die Abschnitte mit dem Einstein-Symbol sind nicht prüfungsrelevant, geben aber für mathematisch Interessierte Hintergrundinformationen, die anderswo wohl nicht zugänglicher zu finden sein werden.

Surveys and other Supplemental Material

Statistik-Formelsammlung (German)

Further Reading

  • Classic Regression Models
    • Jeffrey M. Wooldridge, Introductionary Econometrics, a Modern Approach Thomson South-Western, USA (2002), ISBN 0-324-11364-1.
  • Discrete-Choice Models
    • Ben-Akiva, Lerman, Discrete Choice Analysis: Theory and Application to Travel Demand , MIT Press Series in Transportation Studies, ISBN-10:0262022176). The standard textbook authored by one Nobel-price winner
    • Kenneth Train, Discrete Choice Methods with Simulation. Can be freely downloaded at the Berkeley site. Particularly relevant are the first five chapters.
    • Koppelmann/Bath, a Self Instructing Course in Mode Choice Modeling: Multinomial and Nested Logit Models. Freely downloadable at the University of texas site.

Kommentierte Literaturempfehlungen (deutsch)

  • Lineare Modelle
    • Eckey, Kosfeld, Dreger, Ökonometrie, Gabler-Verlag, Wiesbaden (1995). Behandelt hauptsächlich die "klassische" Ökonometrie, also im Wesentlichen multivariate lineare Modelle, insbesondere Regression. Schöne Darstellung der verschiedenen Modellkategorien der Ökonometrie. Geht gleich im 2. Kapitel zur Vektor-Matrix-Darstellung über. Für die, die es wirklich wissen wollen: Beispielsweise, warum der Fehlerquadratschätzer effizient ist oder warum im Varianzschätzer (n-K-1) und nicht n im Nenner ist.
    • Von Auer, Ökonometrie, Springer-Verlag, Berlin, Heidelberg (1999). Elementarere Darstellung der klassischen linearen Ökonometrie ohne Vektor-Matrix-Notation. Sehr instruktiv ist die kritische Hinterfragung der verschiedenen Annahmen, welche bei der Parameterschätzung linearer Modelle gemacht werden.
    • Joachim Frohn, Grundausbildung in Ökonometrie, de Gruyter, Berlin (1995). Gute Darstellung der Modellspezifikation und der Maximum-Likelihood-Methode.
  • Modelle der diskreten Wahltheorie
    • Gunther Maier und Peter Weiss, Modelle diskreter Entscheidungen, Springer-Verlag Wien New York (1990). Das deutsche Standardwerk für die Modelle der diskrete Wahltheorie. Bemerkenswerterweise wird diese, für di3 Verkehrsökonometrie sehr wichtige Modellklasse in den anderen deutschsprachigen Büchern nahezu komplett ignoriert.
    • W. Schnabel, D. Lohse, Grundlagen der Straßenverkehrstechnik, Band 2 , Verlag für Bauwesen, Berlin, ISBN 3-345-00567-0 Verkehrsplanerischer Hintergrund, nützlich zum Verständnis vieler Aufgaben und Beispiele.
  • Input-Output-Modelle
    • Richter, Verkehrsökonometrie, Oldenbourg-Verlag, München.

Martin Treiber