Venkataraman, V., et al., "Dynamical Regularity for Action Analysis" (2015)

Interesting work by Vinay and Pavan discussing alternative approaches to quantifying dynamical regularity in human movement using approximate entropy methods suggested by Steven M. Pincus in 1991-1995, as an alternative to other parameters, including Lyapunov exponents.  The advantage to ApEn methods is that they require less time-series data.  This could be an interesting approach.

Just spoke with Pavan and he indicated that he would have someone send me some MatLab programs next week for me to review.  Of course, these are batch processes but it may yield something interesting for real-time systems also.

Dynamical Systems shared folder in RefWorks

Hi all,

Experimenting with RefWorks through ASU Library online, and I've created a shared RefWorks folder called Dynamical Systems that I'm hoping you may view.  This would have an ongoing accumulation of reference articles and books relating to Dynamical Systems.  If you don't use RefWorks, then this may be of little value to you.  If you routinely use RefWorks, please take a moment and verify that you can view the shared folder.  There is no link; it should simply appear based on your access of RefWorks through the ASU Library portal (I think).

Thanks,

Brenda

p.s., In any case, I should be able to create a bibliography out of RefWorks that would include all of the items in it.

Some Hadoop notes

  • Hadoop was the name of the toy elephant that the original developer's daughter was playing with when he started the project (based on Google's white paper) ten years ago.
  • The commercial version from Cloudera (and others) allows you to use Amazon servers to run data analysis and queries (AWS = Amazon Web Servers).  Can be very expensive.  Can place a bid for low price for batch projects running off peak time.
  • Another version runs from Microsoft Azure.  Microsoft Azure has a sensor data analysis example.
  • Pig is a query script (power shell) developed at Yahoo.
  • Hadoop is used for behavioral data, not transactional data.  This is how Facebook, for example, knows your cursor position at any point in time and can send you ads even though you haven't actually clicked on something.