Messages
Time: Friday, 26. May, place: room 1020, NHA
09:00 Kirsten Marie Ginnerup
10:00 Heming Smedsrud Aldrin
11:00 P?l Hersleth
13:00 Henrik Wollen
14:00 August Fosse
Time: Tuesday, 30. May, place: room 1020, NHA
09:00 Magdalena Zednickova
10:00 Henrik Lien
11:00 Maximilian Kollnig
12:00 Silje Thorsen
Remember (!) to inform me about the topic of your presentation latest 3 days before your exam!
The (non-digital) oral exam is supposed to be on Friday, 26. May and 30. May, 2023 (starting at 9:00 am), that is
26. May, Friday: 09:00-15:00 (5 students)
and
30. May, Tuesday: 09.00-13:00 (4 students)
Please, let me know (via email) which exam day you prefer!
Place and time will be announced, soon.
The exam procedure is as follows: The exam takes 45 min. and consists of two parts:
1. Talk/presentation of a topic of free choice about extreme value statistics. The title and topic of the presentation are supposed to be communicated to me (by e-mail) latest 3 days (!) before the exam and approved by myself. The length of the talk is limited to 20 minutes and the form of the presentation is up to the candidate (blackboard, beamer slides,...).
2. General questions about extreme value statistics.
The pensum of ...
The exam dates for our oral exam are supposed to be on Friday, 26. May (5 candidates) and Tuesday, 30. May (4 candidates)!
Please, let me know (via email) which exam day you prefer!
Here: MandatoryAssignment
Deadline: Thursday, 20. April, 14:30 (electronic submission via Canvas).
We are supposed to have an oral exam for STK4550!
The date for the oral exam is not fixed yet and will be discussed with the students.
In our last lessons (21. Feb.-15. March) we studied the exact asymptotics of ruin probabilities in the case of large claim sizes, by using e.g. mathematical tools from the theory of regularly varying functions. See Chapter 1 in the book of Embrechts.
Then, we continued with the discussion one of the main results of extreme value theory, that is the theorem of Fisher-Tippett-Gnedenko, which can be regarded as a central limit theorem for maximal claims (Ch. 3 in Embrechts). The latter result is also important for the practical study of large claims, which may cause the spontaneous ruin of an insurance company.
In addition, we have addressed the problem of the characterization of the maximum domain of attraction of extreme value distributions.
Our study plan for the next weeks is the following:
March: Discussion of convergence rates with respect to the Fisher-Tippett-Gnedenko theorem and statistical methods for extre...
Vi startet opp kurset v?rt med en kort innf?ring i ekstremverdistatistikk og teori av store avvik og skaffet oss oversikt over emner vi kommer til ? studere. Dessuten repeterte vi (25./31. jan.) grunnleggende begrep og resultater fra sannsynlighetsregning (f.eks. stokastisk prosess, (betinget) forventningsverdi,...) som vi f.eks. trenger til estimering av ekstremverdi-fordelinger.
Etter vi har studert effektene a