I heartily recommend the notebooks published in this course as excellent applied reference material to estimation and optimization.
I love it how code and coursework are intermingled, reminiscing me of Knuth's Literate Programming [1]
My beef with many other courses offered (including Coursera) is that they use Matlab when it's clearly advantageous to use IPython Notebook as a better experimenting environment. For example, Daphne Koeller's PGM course[2] is still in Matlab and no matter what you do the code looks extremely clumsy and hard to read. N.B. I wrote tens of thousands of lines of Matlab code, including GUI programs, but that does not mean it's a good language to use especially in cases like this.
I took this class when I was in graduate school at Harvard.
The methods introduced in this course can be used in virtually every domain.
I think everyone involved with some sort of data processing should have at least minimal knowledge of Bayesian inference, simulated annealing, data augmentation, sampling, etc...
I bookmarked this for a later date. I am currently taking a Coursera course and one of the projects touched on Monte Carlo methods, they are super cool, I implemented a rudimentary AI that plays tic-tac-toe(well) in like 70 lines of Python.
I love it how code and coursework are intermingled, reminiscing me of Knuth's Literate Programming [1]
My beef with many other courses offered (including Coursera) is that they use Matlab when it's clearly advantageous to use IPython Notebook as a better experimenting environment. For example, Daphne Koeller's PGM course[2] is still in Matlab and no matter what you do the code looks extremely clumsy and hard to read. N.B. I wrote tens of thousands of lines of Matlab code, including GUI programs, but that does not mean it's a good language to use especially in cases like this.
[1] http://en.wikipedia.org/wiki/Literate_programming
[2] https://www.coursera.org/course/pgm