R programming language resources › Forums › Data manipulation › Using R to find events in a data stream
- This topic has 0 replies, 1 voice, and was last updated 13 years, 5 months ago by
JackJ.
- AuthorPosts
- November 13, 2011 at 4:15 pm #449
JackJ
MemberHi,
I new to R so I’ll apologize in advance. The terms I’ll use are likely not what an R programmer would use, but I think you’ll get the idea.
I have a large data set, TheRecording, with with 1 million records (for example) . Each item in the data set has a small (< 10) but arbitrary number of fields. Lets say 4 for example (t,x,y,z). It a time series with 3 values and t is in milliseconds. I have a smaller data set, TheEvent, of length 10, for example. Its the time signature of an event. The x,y,z values could be any value. If this were a 3 channel audio problem, I'd be finding the times that a hand clap occurred. I'd like to find in TheRecording the location of each event that that is substantially similar to TheEvent. I'd like to know how similar the found event was to the Event. I'd like to set a threshold to ignore events that are below the threshold, in the hope that this would speed up the process. My fictional world is a bit more complex in that the hand claps I'm looking for could be slow (i.e greater than 10 records long) or fast (shorter than 10 records). Is R a good choice for solving this kind of problem? If so can you point towards a place to start. If not, any ideas? Thanks
- AuthorPosts
- You must be logged in to reply to this topic.