Till KTH:s startsida Till KTH:s startsida

Assignment 1

This is the first assignment. You are supposed to individually and without consulting anything else than your book solve the following excersizes from it: 2.8, 3.7, 3.17, and 5.8. The assignment is due 13.15 Friday Nov. 28, 2014. This is a strict deadline. Send your solutions to me (Jens Lagergren <jensl@csc.kth.se>) and Kristoffer (Kristoffer Sahlin <kristoffer.sahlin@scilifelab.se>).

Lärare Jens Lagergren skapade sidan 13 november 2014

kommenterade 14 november 2014

I am not even allowed to consult BETA?

Lärare kommenterade 19 november 2014

You are allowed to use BETA.

kommenterade 27 november 2014

Why are we only allowed to consult the book for information? I find it really hard answering the questions just from reading the related chapters, and I generally use multiple sources when trying to understand new concepts.

Lärare kommenterade 27 november 2014

Well sure. What I mean is basically that you should decide on your course material. Learn about the content in the chapters we cover, using that material. Then when doing the assignments you stick to your material. In particular, you don't start surfing the web in order to find something that appears to be related to the question. 

kommenterade 27 november 2014

There is a maximum of eight points on this assignment right? Is it two points per exercise?

Lärare kommenterade 27 november 2014

Sorry that is old info. I'll make sure that you get new info asap. 

kommenterade 1 december 2014

When will the results of assignment 1 be published in rapp? 

Lärare kommenterade 2 december 2014

Kristoffer has been in the US, but we will have a meeting this week.

Lärare Jens Lagergren ändrade rättigheterna 3 december 2014

Kan därmed läsas av alla och ändras av lärare.
kommenterade 7 december 2014

Will the points for assignment 1 be published anywhere? I would like to know how much I got. 

Lärare kommenterade 8 december 2014

Kristoffer estimates that it will be done by Wednesday. We have asked those responsible for Rapp to provide the necessary means to communicate the results through Rapp (so hopefully they will also be done by then). 

Lärare kommenterade 9 december 2014

Notice my comment above concerning weights. 

kommenterade 10 december 2014

Assignment 1 has been corrected and reported in RAPP. If you submitted solutions but you cant see the grading, I have probably not gotten them. Exeptions are Ahmed Laeeq, Johan Markdahl, David Menendez, Ayesha Tasnim and Daan Wynen. I have corrected you assignments but you are not in rapp under "statmet14" so your result have not been reported. Jens, can you add them to the course?

My notes during grading:

*Mostly trouble with 2.8 and 5.8.

* 2.8: iff requires showing both => and <=.

*3.7: Think of data as a set of observations. Lots of solutions treat the data (x,d, or whatever notation) as a single value which inhibits the conclusions one can make from the assignment, i.e. that the MAP and ML estimates agree for alpha,beta=>0.  This becomes more clear for n>1 observations, otherwise the ML and MAP is just x/1 (d/1), i.e. the data point itself.

*5.8: Use log likelihood when deriving and calculating ML estimates, lots of miscalculations has lead to misinterpretations, I have been fairly generous though.