STK4110 - Spring 2011

Teaching plan

DateTeacherPlaceTopicLecture notes / comments
21.01.2011Bent Natvig  B81 N.H. Abels hus  Introduction and motivating examples on the use of the theory.  As part of the introduction Bayes formula and Bayes theorem will be recalled as a basis for highlighting parts of the following paper, which will be handed out,

Meinhold, R. J. and Singpurwalla, N.D. (1983) Understanding the Kalman Filter, The American Statistician, 37, 123-127.

The first chapter of West and Harrison (1997) should be read on your own. 

28.01.2011Bent Natvig  B81 N.H. Abels hus  Traditional time series analysis.  Main points from West and Harrison (1997), Chapter 9.4, pages 291- 301 were covered. 
04.02.2011Bent Natvig  B81 N.H. Abels hus  Ideas of probability and Bayesian forecasting and the first order polynomial DLM.  A part of Jeff Harrison`s lecture series on the subject given at the University of Warwick in the autumn of 1994 was presented. Then we started lecturing on Chapter 2 in West and Harrison (1997) treating the first order polynomial Dynamic Linear Model (DLM). The material until the end of the first proof of Theorem 2.1 was covered. The second proof can be dropped. 
11.02.2011Bent Natvig  B81 N.H. Abels hus  The first order polynomial DLM.  We covered until 2.4 in West and Harrison (1997). Hopefully part of the lectures was clearer than the book. Example 2.1, 2.3.2 and 2.3.4 should be read on your own. The last paragraph of 2.3.3 and the whole of 2.3.6 can be dropped. 
18.02.2011Bent Natvig  B81 N.H. Abels hus  The first order polynomial DLM.  We covered 2. 4 and discussed the exercises 1,4,6,7. 
25.02.2011Bent Natvig  B81 N.H. Abels hus  The first order polynomial DLM.  We covered background properties of the normal, gamma and T-distributions to be used in 2.5. and this section until the end of the proof of Theorem 2.4. We discussed the exercises 5 and 10 a). 
04.03.2011Bent Natvig  B81 N.H. Abels hus  The dynamic regression model  Some further comments to Theorem 2.4 and Section 2.6 was given. The latter section should be read on your own. Then some main parts of Chapter 3 was covered, again this chapter should be read on your own. You can skip the proof of Theorem 3.1 since this theorem is a special case of Theorem 4.1 to be proved next week. We discussed exercise 17 a)-d). 
11.03.2011Bent Natvig  B81 N.H. Abels hus  The dynamic linear model  We covered Chapter 4 until the end of the proof of Theorem 4.1. We discussed exercise 3 and most of exercise 6 from Chapter 3. Note that there is a mistake in the second to last line of exercise 6. 
18.03.2011Bent Natvig  B81 N.H. Abels hus  The dynamic linear model  We covered the rest of Section 4.3, Section 4.4, commented on Sections 4.5 and 4.6 and then treated the second order trend model in 7.2.1. We discussed the rest of exercise 6 and exercise 7 of Chapter 3 in addition to exercise 6 of Chapter 4. 
25.03.2011Bent Natvig  B81 N.H. Abels hus  Case study from insurance  We started by covering some highlights from Section 8.6 Fourier form representation of seasonality to have the background for discussing the paper, which has been handed out:

Tvete, I.F. and Natvig, B. (2002) A Comparison of an Analytical Approach and a Standard Simulation Approach in Bayesian Forecasting Applied to Monthly Data from Insurance of Companies. Methodology and Computing in Applied Probability, 4, 95-113.

This was covered until the middle of page 111. Ideally, this should be looked into by you beforehand. 

01.04.2010Bent Natvig  B81 N.H. Abels hus  Case study from earthquake data modeling  We started by finishing the discussion on the paper Tvete and Natvig (2002). We then proceeded to discuss the paper, which has been handed out:

Natvig, B. and Tvete, I.F. (2007) Bayesian Hierarchical Space- time Modeling of Earthquake Data. Methodology and Computing in Applied Probability, 9, 89-114.

This was covered until the end of Table 1, page 99. Ideally, this should be looked into by you beforehand. 

08.04.2011Bent Natvig  B81 N.H. Abels hus  Case study from earthquake data modeling  We started by finishing the discussion on the paper Natvig and Tvete (2007). We then covered Section 4.7 Filtering recurrences until the beginning of Theorem 4.4, in addition to the material on conditional distributions in 17.2.2. 
15.04.2011Bent Natvig  B81 N.H. Abels hus  Filtering recurrences  We finished Section 4.7 and then proceeded to cover Section 4.9 Linear Bayes` optimality until the middle of the proof of Theorem 4.9.  
29.04.2011Bent Natvig  B81 N.H. Abels hus  Linear Bayes`optimality  We finished Section 4.9.2, 4.9.3 and 4.9.5. Then we covered Example 9.5 in Section 9.4.6 ARMA models in DLM form. 
06.05.2011Bent Natvig  B81 N.H. Abels hus  Model monitoring  We start by covering Examples 9.6 and 9.7 in Section 9.4.6 ARMA models in DLM form. We will then cover Section 11.4.1 Bayes` factors for model assessment, Section 11.4.2 Cumulative Bayes` factors and Example 11.3 from Section 11.4.3 Specific alternatives for the DLM. 
Published Nov. 25, 2010 11:04 AM - Last modified Apr. 29, 2011 6:19 PM
E-mail this page