iSAX 2.0: Indexing and Mining One Billion Time Series [pdf]

cs.ucr.edu

26 points·chaotic-good··8 comments

Open articleView on HN

This iSAX 2.0 paper seems to be from 2010, so the HN title should probably be changed.

The first iSAX paper was from 2008, and there's a page here about it:

http://www.cs.ucr.edu/~eamonn/iSAX/iSAX.html

The list of "selected" publications by Eamonn Keogh is a whole lot of fun:

http://www.cs.ucr.edu/~eamonn/selected_publications.htm

iSAX was an improvement over simple SAX indexing scheme that was known since 2005 or so.

My failing meatware recollection said 2001-ish, maybe earlier, but it seems I'm wrong according to this:

http://www.cs.ucr.edu/~eamonn/SAX.htm

>"SAX was invented by Eamonn Keogh and Jessica Lin in 2002, using funding from NSF Career Award 0237918."

Does anyone know of any code for iSAX that isn't in C# or Java? I'd quite like to get stuck in but right now I'm using Python. I've got SAX set up but getting the iSAX indexing will take some time.

Both of these appear to do what you want:

- https://github.com/dolaameng/pysax - https://github.com/nphoff/saxpy

Thank you. These are great for SAX but unfortunately don't implement iSAX. I'm working on iSAX for Python and will release that once it is complete.

Interesting coincidence. We just went over this paper last month in our white paper group.

:)) great Themis! super teacher

Reply on news.ycombinator.com