Wednesday, November 28, 2012

A Stochastic Correlations Model for Default Risk

by Toby Daglish¤and Shaloub Razak
http://www.nzfc.ac.nz/archives/2008/papers/0820.pdf

interesting model for default probabilities used in a cdo: closed form for single firm, monte carlo for multiple firms. calibrate with credit default swap using an extended kalman filter. explains changing default correlation over time in copula models.

empirical mode decomposition

interesting work by danilo mandic at imperial college on emd and related stuff:
http://www.commsp.ee.ic.ac.uk/~mandic/

time-frequency analysis for multichannel and nonstationary data, specifically with emd:
http://www.commsp.ee.ic.ac.uk/~mandic/research/emd.htm

some really impressive stuff with frequency and mode separation, much better than ordinary wavelets.

Multivariate multiscale entropy: A tool for complexity analysis of multichannel data

http://www.commsp.ee.ic.ac.uk/~mandic/research/MUA_DPM_Multivariate_MSE_PRE_2011.pdf

just skimmed an interesting article about estimating multivariate entropy, over multiple time scales. possible applications for mutual information estimation?

Saturday, November 24, 2012

texworks and shell escape

tried to get shell escape working with texworks, so i could do naughty and dangerous things with write18 from latex. there are checkboxes that sound like they are made for this sort of thing in the preferences, but i found them to have no such effect. finally edited the options for the typesetting commands for pdflatex, and added --shell-escape. now it works. unfortunately, i couldn't get it work for the texify combo (pdflatex, makeindex, bibtex) but whatever.

oddly, the same option works for texworks with either miktex or texlive, even though they are different on the command line: -shell-escape for texlive and -enable-write18 for miktex.

Saturday, November 17, 2012

sobol sequences and python

tried it, seems to work just fine.

https://github.com/naught101/sobol_seq/blob/master/sobol_test_output.txt
http://people.sc.fsu.edu/~jburkardt/py_src/sobol/sobol.html

latex building

here are some tools for building documents from latex/tex.

texcaller has interfaces for python and c++, among others:
http://www.profv.de/texcaller/download.html

http://obrecht.fr/texwrapper/
http://latex-mk.sourceforge.net/
http://mirror.ox.ac.uk/sites/ctan.org/support/latexmk/latexmk.pdf
http://tug.ctan.org/pkg/autolatex

python + latex

python.sty looks very good, very small, relatively simple, and almost everything i want for calling python from latex. only problem is that it starts a new interpreter for each python environment block.

https://bitbucket.org/brotchie/python-sty/changesets/tip/branch(%22default%22)

to use python.sty with rubber, i will likely need this:
http://bazaar.launchpad.net/~brotchie/rubber/shell-escape/revision/419?start_revid=419

here's some good info on working with files with tex/latex:
http://stackoverflow.com/questions/2115379/write-and-read-from-a-latex-temporary-file

a latex builder similar to rubber, but maybe simpler and easier. i'd reffed this before:
https://bitbucket.org/nmandery/python-pdflatex/

more info on calling out to python from latex:
http://thewikiblog.appspot.com/blog/python-inside-latex
https://github.com/gpoore/pythontex/tree/master/pythontex

mahotas

http://mdp-toolkit.sourceforge.net/

interesting package that's apparently well known in the computer vision world. python, with numerical stuff in c++. algorithms include SURF, watershed, thresholding, convex hull, polygon drawing, haralick textures, local binary patterns, zernike moment, distance transform, and freeimage and imread interface.

modular toolkit for data processing

http://mdp-toolkit.sourceforge.net/

interesting project with a number of capabilities. python code for pca, ica, slow feature analysis, manifold learning methods ([Hessian] local linear embedding), classifiers, factor analysis, rbm, etc.

according to the 'intro to scipy' talk at pydata 2012, it has the fastest pca available in python (even if the interface is more difficult than scipy svd or sklearn.decomposition.PCA).

numba and cython

interesting comparison between numba and cython (and pure python). both projects i want to keep and eye on.

http://jakevdp.github.com/blog/2012/08/24/numba-vs-cython/