Topics in Info Theory
From enfascination
| (One intermediate revision by one user not shown) | |||
| Line 1: | Line 1: | ||
==Fall 09== | ==Fall 09== | ||
===social work=== | ===social work=== | ||
| − | application of MI to groups of people [http://dimacs.rutgers.edu/Workshops/Defense/abstracts.html] | + | application of MI to groups of people [http://dimacs.rutgers.edu/Workshops/Defense/abstracts.html] and [http://scholar.google.com/scholar?q=adibi+link+discover] |
===causality and direction=== | ===causality and direction=== | ||
[[Causality detection based on information-theoretic approaches in time series analysis Hlavackova-Shindler Palus 2009| review]]<bibtex>@article{hlaváčková2007causality, | [[Causality detection based on information-theoretic approaches in time series analysis Hlavackova-Shindler Palus 2009| review]]<bibtex>@article{hlaváčková2007causality, | ||
| Line 14: | Line 14: | ||
} | } | ||
</bibtex> | </bibtex> | ||
| + | ===disembodied notes from Olaf=== | ||
| + | *resampling to uniform or by rank | ||
| + | *rule of thumb for data requirements is (3 to 5)*n^m time steps are standard for significance, where n is the number of bins per variable and m is the number of variables. This suggests that 1 2 and maybe maybe 3 variables are tractable and beyond that it gets ugly. | ||
| + | *there is also some other heuristic about every possible bin having atleast 3 data points. | ||
| + | *all heuristics should be given credit commensurate with their rigor. | ||
==Fall 08== | ==Fall 08== | ||
Latest revision as of 02:50, 13 November 2009
Contents |
Fall 09
social work
application of MI to groups of people [1] and [2]
causality and direction
review<bibtex>@article{hlaváčková2007causality,
title=Template:Causality detection based on information-theoretic approaches in time series analysis, author={Hlav{\'a}{\v{c}}kov{\'a}-Schindler, K. and Palu{\v{s}}, M. and Vejmelka, M. and Bhattacharya, J.}, journal={Physics Reports}, volume={441}, number={1}, pages={1--46}, year={2007}, publisher={Elsevier}
} </bibtex>
disembodied notes from Olaf
- resampling to uniform or by rank
- rule of thumb for data requirements is (3 to 5)*n^m time steps are standard for significance, where n is the number of bins per variable and m is the number of variables. This suggests that 1 2 and maybe maybe 3 variables are tractable and beyond that it gets ugly.
- there is also some other heuristic about every possible bin having atleast 3 data points.
- all heuristics should be given credit commensurate with their rigor.
Fall 08
The first Google hit for "course graining in information theory"
" IGP Grain Purchasing Short Course Syllabus Makes the international grain buyer aware of the available information provided by ... Explains the theory of futures and options by illustrating the use of ... www.k-state.edu/igp/grainsyllabus.htm - 14k - Cached - Similar pages - Note this "
Google this too: "coarse-grain to fine-grain parallelism"
See this for more questions: