There will be one exercise in each part. You have time to work on these during two weeks after the lectures on each part. You will choose one of the exercises and re-deliver before the exam. This will count 30 % of the grade.
We encourage you to work in pairs, and one copy is to be handed in for each pair.
For coding, you may choose from matlab, python, R and C/C++, my preference would be matlab. If you do your code in C/C++ it needs to have extremely detailed comments. matlab and R are probably the easiest to use, as they have a lot of inbuilt functions that need not be programmed. If you are unsure, however, of how these are implemented, I will encourage to do an implementation yourselves.
It is probably wise to look at the exercises well ahead of the delivery date, and start programming as soon as possible.
[15/3] Problem 3 in Exercise 2 will not be mandatory, but you are welcome to try it if you want. [7/4] I have noticed that Exercise 3 is not possible to download on Windows machine due to some metadata. I will try to fix this, but it should be possible to use one of the machines on the third floor and print it there.
[18/4] I believe I have scored everybody's second assignment. E-mail me if I have forgotten you.
- Exercise 1: Continuous random fields: Exercise 1 (Deadline Friday 25 feb, 16.00). Data set: Topo, a matlab-friendly version: topo.mat. A kriging toolbox for matlab that is useful to look at can be found at mGstat. Note that there is a typo in the exercise. You should do Kriging in the square [0,6.5]x[0,6.5], not [0,350]x[0,350]. I will upload a corrected version as soon as possible.
Erlend Aune is responsible for the exercises. Contact: erlenda (at) math.ntnu.no