I want to make a cognitive radio. I read Essentials of Cognitive Radio: quite a nice book to begin with, I now know CR's architecture and I'm going to build my first cognitive engine.
Yet, I'm confused by the quantity of information: I'm a nearly graduated student but they never taught us the basis of machine learning (even if we studied mathematical analysis and statistics).
From the book mentioned above, I summarized the following concept:
A cognitive engine can optimize, reason, learn, decide.
Later in the book, four big topics are introduced:
- knowledge representation;
- optimisation techniques;
- reasoning methods;
- machine learning approaches.
Then, the book asks the reader a question: "what algorithm(s) will be used for each topic?"
Nice question: will you gentlemen share your knowledge about this matter? Do you have some experience in Cognitive Radio's field? Were some algorithms better than other?
I know the answer could be "it depends on your needs" and this is right: but I'd prefer to hear some opinions first and have at least 2 algorithms for each topic, before starting to work.