# Which sections to cover for my graduation thesis on blind source separation?

I'm not sure if I'm on right place to ask a simple suggestion, so please be kind and don't downvote my question. I'm not asking someone to do it for me, I just need short guidance how should I proceed.

Currently, I'm writing a study of Blind source separation using eigenvalue decomposition. Since I'm not good at writing, I have done so far the needed programming part in Matlab. The study must be composed of ~50 pages theory about the problem. Theory should be in terms of explaining the BSS, BSS methods, ICA, and EVD which is the main topic. Can you suggest what to include in my writing since BSS has a lot of methods and it is difficult for me to decide and follow what to write and what not write. Can someone outline the order of things I should write and mention in my study?

I'm stuck at what should I write next, so far I have:

Introduction to BSS
Definition of Independence of signals
Independence and correlation
PCA and other prewhitening methods:
1. centering
2. whitening


I'm using books like: Independent Component Analysis - A Tutorial Introduction Handbook of Blind Source Separation Independent Component Analysis and Applications 2010 - Pierre Comon, Christian Jutten

Thank you,

• Shouldn't you focus on why your implementation is better than all the rest, or the specific contexts in which it is better and in which it is worse? Does it deal with delay between the channels, as in real recordings from multiple microphones? Can it handle echoes, etc – endolith Aug 13 '13 at 14:20

I would structure the thesis like this:

Introduction/Problem Statement (i.e. whatever problem BSS solves)
How BSS solves that problem (high level)
Theory behind BSS
1) Definition of Independence of signals
2) Independence and correlation
3) PCA and other prewhitening methods:
1. centering
2. whitening
1) How those results compare to what the theory says should be achievable
Possible Improvements
Conclusion


The "Possible Improvements" section should have ideas on how you or the reader could take what you have done and improve upon it, by (for example) improving the implementation, testing it in new ways, or applying the implementation to different problems.

The Conclusion should summarize the paper, and in particular the problem statement, how your implementation solves that problem, and what results you achieved.

You need to start as broad as possible and work your way towards your practical work. Your writing should make the case for the practical EVD work that you've done in matlab.

After the start you posted you're going to need to need to cover bits about optimisation algorithms as well. FastICA, Infomax etc. Also different measures of independence: entropy, kurtosis etc.

My starting point for ICA work is Naik & Kumar 2011

I've taken this course. I think using these two papers plus those books you mentioned would be so useful in writing a report, thesis , ... .

Blind signal separation: statistical principles

Independent Component Analysis (A Tutorial)