alright this is what you want to do.
install.packages("fields", dependencies=TRUE)
tim.colors is in this package and it has a blue to red color scheme-
blue being the lowest and red being the highest. This color scheme
makes sense to me and is a common thing that a people (read engineers)
familar with matlab or the like will understand.
USE the morlet wavelet it is compactly supported which means that it
quickly goes to zero once it gets out of the scale that it is fitting.
Making it good for a localized fit.
what you are looking at is the modulus (absolute value) of the
convolution of the wavelet with the signal at a particular scale (kind
of like frequency in fourier analysis) on the y-axis through time
(local fitting) on the x-axis. Your are trying to find periodicity?
I kind of think of wavelet analysis as the partitioning of variance of
the signal into continuous scale.
because of algorithm calculation the scale is in log2(value of the
time series) so to get to your time units (which you set in the deltat
or frequency argument when you create a timeseries with ts() )
2^(value of the scale).
I hope this helps
Stephen
2008/9/21 yuankun shi <shiyuankun.debian at gmail.com>:
Thanks, I have succeeded to do this, first wavCWTPeaks to get every peaks' coordinate, then calculated their horizontal distance, finally,bkde output the distance's distribution, that's what I want. On the contrary, picture of wavCWT seems hard to understand, I am not sure what the y axis and the color mean. Could you do me a favor? 2008/9/19 stephen sefick <ssefick at gmail.com>
I would suggest wavelet analysis- library wmtsa wavCWT This will tell you if there is a periodicity localized in time which fourier analysis canno tell you- if the variance is not constant through time then you should use this. 2008/9/19 yuankun shi <shiyuankun.debian at gmail.com>:
I have spent lots of time to download the code you have mentioned. But all of them is not I wanted, except the latest one, I have not found it anywhere. Maybe I have not make my problem clearly, sorry for that. I have a series data, it consists of time and rate. To plot rate vs time in picture, I found it has perodicity to some extent. The rate rise and fall with time, but not with fixed cycle and fixed amplitude. So I am wondering, is there any tools to get the cycle? and furthmore, to draw it's density picture? Since there is bkde in package KernSmooth, so the 2nd is not strict needed 2008/9/11 stephen sefick <ssefick at gmail.com>
all of the functions that I listed are time series tools for looking at what I think you want. this can be done you just have to understand the methodology. So, look at some of the things that I suggested, If these don't help then I don't understand what you want, and it is necissary for you to help me figure out what it is that you want. good luck 2008/9/11 yk <shiyuankun.debian at gmail.com>:
The data I mentioned above is oscilating vs time?but there are not obersevable fixed cycle if I just plot this data. How to get the average cycle?or the most probable range of cycle with statistical methods? I don't know how to achieve it by R, is there any command? On Sep 11, 10:52 am, "stephen sefick" <ssef... at gmail.com> wrote:
?spectrum ?acf ?ccf library(wmtsa) ?wavCWT library(sowas) ?wsp you could also look at lagged plots to look for periodicity. if you elaborate on the problem and include executable sample code you will probably recieve more help. On Wed, Sep 10, 2008 at 10:02 PM, yk <shiyuankun.deb... at gmail.com> wrote:
There is a series of data contains time in fixed step and energy varying with time, how to test its periodicity?In R, it seems there is no direct tools since I have search the R manual with periodic and I have not found any related topic. Thanks a lot
______________________________________________ R-h... at r-project.org mailing list https://stat.ethz.ch/mailman/listinfo/r-help PLEASE do read the posting guidehttp://www.R-project.org/posting-guide.html and provide commented, minimal, self-contained, reproducible code.
--
Stephen Sefick
Research Scientist
Southeastern Natural Sciences Academy
Let's not spend our time and resources thinking about things that
are
so little or so large that all they really do for us is puff us up
and
make us feel like gods. We are mammals, and have not exhausted the
annoying little problems of being mammals.
-K. Mullis
______________________________________________ R-h... at r-project.org mailing listhttps://stat.ethz.ch/mailman/listinfo/r-help PLEASE do read the posting guidehttp://www.R-project.org/posting-guide.html and provide commented, minimal, self-contained, reproducible code.
______________________________________________ R-help at r-project.org mailing list https://stat.ethz.ch/mailman/listinfo/r-help PLEASE do read the posting guide http://www.R-project.org/posting-guide.html and provide commented, minimal, self-contained, reproducible code.
--
Stephen Sefick
Research Scientist
Southeastern Natural Sciences Academy
Let's not spend our time and resources thinking about things that are
so little or so large that all they really do for us is puff us up and
make us feel like gods. We are mammals, and have not exhausted the
annoying little problems of being mammals.
-K. Mullis
--
Stephen Sefick
Research Scientist
Southeastern Natural Sciences Academy
Let's not spend our time and resources thinking about things that are
so little or so large that all they really do for us is puff us up and
make us feel like gods. We are mammals, and have not exhausted the
annoying little problems of being mammals.
-K. Mullis
Stephen Sefick Research Scientist Southeastern Natural Sciences Academy Let's not spend our time and resources thinking about things that are so little or so large that all they really do for us is puff us up and make us feel like gods. We are mammals, and have not exhausted the annoying little problems of being mammals. -K. Mullis