Universität Wien
Achtung! Das Lehrangebot ist noch nicht vollständig und wird bis Semesterbeginn laufend ergänzt.

040025 UK Large-Scale Inference (2018S)

8.00 ECTS (4.00 SWS), SPL 4 - Wirtschaftswissenschaften
Prüfungsimmanente Lehrveranstaltung
Moodle

An/Abmeldung

Hinweis: Ihr Anmeldezeitpunkt innerhalb der Frist hat keine Auswirkungen auf die Platzvergabe (kein "first come, first served").

Details

max. 30 Teilnehmer*innen
Sprache: Englisch

Lehrende

Termine (iCal) - nächster Termin ist mit N markiert

  • Donnerstag 01.03. 13:15 - 14:45 Seminarraum 3 Oskar-Morgenstern-Platz 1 1.Stock
  • Dienstag 06.03. 11:30 - 13:00 Seminarraum 3 Oskar-Morgenstern-Platz 1 1.Stock
  • Donnerstag 08.03. 13:15 - 14:45 Seminarraum 3 Oskar-Morgenstern-Platz 1 1.Stock
  • Dienstag 13.03. 11:30 - 13:00 Seminarraum 3 Oskar-Morgenstern-Platz 1 1.Stock
  • Donnerstag 15.03. 13:15 - 14:45 Seminarraum 3 Oskar-Morgenstern-Platz 1 1.Stock
  • Dienstag 20.03. 11:30 - 13:00 Seminarraum 3 Oskar-Morgenstern-Platz 1 1.Stock
  • Donnerstag 22.03. 13:15 - 14:45 Seminarraum 3 Oskar-Morgenstern-Platz 1 1.Stock
  • Dienstag 10.04. 11:30 - 13:00 Seminarraum 3 Oskar-Morgenstern-Platz 1 1.Stock
  • Donnerstag 12.04. 13:15 - 14:45 Seminarraum 3 Oskar-Morgenstern-Platz 1 1.Stock
  • Dienstag 17.04. 11:30 - 13:00 Seminarraum 3 Oskar-Morgenstern-Platz 1 1.Stock
  • Donnerstag 19.04. 13:15 - 14:45 Seminarraum 3 Oskar-Morgenstern-Platz 1 1.Stock
  • Dienstag 24.04. 11:30 - 13:00 Seminarraum 3 Oskar-Morgenstern-Platz 1 1.Stock
  • Donnerstag 26.04. 13:15 - 14:45 Seminarraum 3 Oskar-Morgenstern-Platz 1 1.Stock
  • Mittwoch 02.05. 15:00 - 16:30 Seminarraum 6 Oskar-Morgenstern-Platz 1 1.Stock
  • Donnerstag 03.05. 13:15 - 14:45 Seminarraum 3 Oskar-Morgenstern-Platz 1 1.Stock
  • Dienstag 08.05. 11:30 - 13:00 Seminarraum 3 Oskar-Morgenstern-Platz 1 1.Stock
  • Dienstag 15.05. 11:30 - 13:00 Seminarraum 3 Oskar-Morgenstern-Platz 1 1.Stock
  • Donnerstag 17.05. 13:15 - 14:45 Seminarraum 3 Oskar-Morgenstern-Platz 1 1.Stock
  • Mittwoch 23.05. 15:00 - 16:30 Seminarraum 3 Oskar-Morgenstern-Platz 1 1.Stock
  • Donnerstag 24.05. 13:15 - 14:45 Seminarraum 3 Oskar-Morgenstern-Platz 1 1.Stock
  • Dienstag 29.05. 11:30 - 13:00 Seminarraum 3 Oskar-Morgenstern-Platz 1 1.Stock
  • Dienstag 05.06. 11:30 - 13:00 Seminarraum 3 Oskar-Morgenstern-Platz 1 1.Stock
  • Donnerstag 07.06. 13:15 - 14:45 Seminarraum 3 Oskar-Morgenstern-Platz 1 1.Stock
  • Dienstag 12.06. 11:30 - 13:00 Seminarraum 3 Oskar-Morgenstern-Platz 1 1.Stock
  • Donnerstag 14.06. 13:15 - 14:45 Seminarraum 3 Oskar-Morgenstern-Platz 1 1.Stock
  • Dienstag 19.06. 11:30 - 13:00 Seminarraum 3 Oskar-Morgenstern-Platz 1 1.Stock
  • Donnerstag 21.06. 13:15 - 14:45 Seminarraum 3 Oskar-Morgenstern-Platz 1 1.Stock
  • Dienstag 26.06. 11:30 - 13:00 Seminarraum 3 Oskar-Morgenstern-Platz 1 1.Stock
  • Donnerstag 28.06. 13:15 - 14:45 Seminarraum 3 Oskar-Morgenstern-Platz 1 1.Stock

Information

Ziele, Inhalte und Methode der Lehrveranstaltung

The term ``big data" is bandied about so frequently that it has lost perhaps any relevant meaning, instead merely referring to a vaguely connected set of ideas related to ``modern data sets." While data set size does pose computational problems, our goal will be understanding some of the corresponding changes to statistical methodology. This transition is summarized best in the Prologue of Brad Efron's book, Large Scale Inference, and was the inspiration for the title of this course:

``At the risk of drastic oversimplification, the history of statistics as a recognized discipline can be divided into three eras:
1. The age of Quetelet and his successors, in which huge census-level data sets were brought to bear on simple but important questions: Are there more male than female births? Is the rate of insanity rising?
2. The classical period of Pearson, Fisher, Neyman, Hotelling, and their successors, intellectual giants who developed a theory of optimal inference capable of wringing every drop of information out of a scientific experiment. The questions dealt with still tended to be simple—Is treatment A better than treatment B? — but the new methods were suited to the kinds of small data sets individual scientists might collect.
3. The era of scientific mass production, in which new technologies typified by the microarray allow a single team of scientists to produce data sets of a size Quetelet would envy. But now the flood of data is accompanied by a deluge of questions, perhaps thousands of estimates or hypothesis tests that the statistician is charged with answering together; not at all what the classical masters had in mind."

Clearly we will be addressing section 3. Topics covered include:
- Testing problems in high dimensions: sparse and nonsparse alternatives.
- Multiple testing problems: familywise error rate (FWER), closure-principle, procedures for controlling FWER, false discovery rate (FDR), procedures for controlling FDR, empirical Bayes interpretation of FDR.
- Model selection in high dimensions: thresholding rules, Lasso, Dantzig.
- Post-selection inference: POSI, Selective inference, Knockoffs, multiple comparisons.
- James-Stein estimation, Stein's unbiased risk estimate, empirical Bayes view of James-Stein Prediction error.

Art der Leistungskontrolle und erlaubte Hilfsmittel

Homework, Final, Project, Participation. Of the 4, the project will be given the highest weight. The final will be largely conceptual. Subject to change.

Mindestanforderungen und Beurteilungsmaßstab

In preparation for the course, I recommend revising the following chapters from Keener 2010, Theoretical Statistics: Topics for a Core Course; 1-4, 6-8, 12, 14. You may skip the optional sections. If a section is not review, please let me know.

Prüfungsstoff

Literatur

Zuordnung im Vorlesungsverzeichnis

Letzte Änderung: Mo 07.09.2020 15:28