Information and coding theory (20162017)
Last modification: 28 May 2017 10:00
Informations
 In the statement of the first TP, exercise 7 is now: Prove that H(X ,YZ) = H(X Z)+ H(YX,Z). The previous expression was wrong.
 In the project 3, be careful that data matrices have been transposed! In order to use your previous codes, you only have to transpose matrices.
 Do not forget to read and follow the policy for plagiarism (here) otherwise, you expose yourself to measures taken by the ULg against plagiarism (here).
 New with MatLab? I suggest you to follow this short tutorial to get used to Matlab (here).
 Follow the guidelines to submit your project. The deadline is the deadline: delays are not acceptable. The mail subject and the name of the archive have to be correct.
 Bayes Net Toolbox for Matlab (required for all projects): Download, Installation and Documentation.
Agenda
07/02  Theoretical course Presentation of the practical part of the course 

14/02  Theoretical course Project 0 

20/02  Deadline  Deadline for project 0 
21/02  Theoretical course Exercise session : Entropy calculus (Statement) 
To prepare: Revise lesson 1 and lesson 2. 
02/03  Project 1  Statement 
07/03  Q&A  About project 1 
14/03  Q&A  About project 1 
15/03  Deadline  Deadline for project 1 
21/03  Project 2 Exercise session : Graphical models (Statement) 
Statement and data 
9/04  Deadline  Deadline for project 2 
18/04  Project 3  Statement and data 
9/05  Deadline  Deadline for project 3 
Frequently asked questions (F.A.Q.)
It is strongly recommanded to participate to Q&A sessions and to read the F.A.Q. section before any mail question.
Project 1
 Let us assume that you have an array TAB (as generated by generate_samples for example) with 3 variables/columns (var 1 = A, var 2 = B, var 3 = C) and 100 lines/samples, entropy(1,TAB) will give you H(A) and cond_entropy(3,2,TAB) will give you H(CB) where "var" and "cond" are variable id or simply the number of the column corresponding to the variable.
Moreover, for question 7 , you could adapt your function to deal with cond_entropy(3,[1,2],TAB) which will return H(CA,B). You can also "modify" the array "samples" to aggregate several columns into one ([0 1 6] > '016' for example) and still use the function of question 2.  Be careful, the corresponding number to each variable can be retrieved in the given script part2.m. Using the order in the statement by counting is not correct.
 Of course, "PRESS" is equivalent to "PRSS".
 In the statement of project 2, the justification of Q10 is related to Q7 and not Q6.
Project 2