Hi all,

an update about the situation of LombScargle.jl
<https://github.com/giordano/LombScargle.jl>

2016-07-18 10:15 GMT+02:00 Mosè Giordano:

> In the future I may implement another much-faster Lomb-Scargle algorithm
> by Press & Rybicki (1989, ApJ, 338, 277), which however requires the data
> to be equally sampled (but in this case also the FFT can be used).
>

I managed to implement this method, so now users can choose between three
different implementations of the Lomb-Scargle periodogram:

   - the original algorithm, that doesn't deal with uncertainties in the
   data nor for a non-null mean of the signal (based on Townsend, 2010, ApJS,
   191, 247)
   - the so called generalised Lomb-Scargle algorithm, that takes into
   account uncertainties and a non-zero mean of the signal (based on
   Zechmeister, Kürster, 2009, A&A, 496, 577)
   - an approximation of the generalised Lomb-Scargle algorithm, made very
   fast (and with lower computational complexity than the true Lomb-Scargle
   algorithm) by some tricks suggested by Press & Rybicki, 1989, ApJ, 338,
   277.  The only downside is that this method requires the data to be equally
   sampled.  In the package, this is controlled by the type of the time
   vector: if it's a Range, this fast method is used (but can be opted-out
   with a keyword)

Other notable features introduced since the first release:

   - you can choose between 7 different normalizations
   - functions to compute the false-alarm probability
   - a function, findmaxpower, to quickly find the maximum power value of
   the periodogram
   - in findmaxfreq it is now possible to restrict the search for the
   frequency with the highest power to a specific range, instead of using the
   whole periodogram
   - a new function, LombScargle.model, to get the Lomb-Scargle model that
   best fits your data at a given frequency (without performing an actual
   periodogram, because this is done just for one frequency)

The complete manual, with some examples complemented with pictures, is
available at http://lombscarglejl.readthedocs.io/ (here
<http://readthedocs.org/projects/lombscarglejl/downloads/pdf/latest/> the
PDF version).

The latest version of LombScargle.jl
<https://github.com/giordano/LombScargle.jl> is v0.1.1, be sure to update
your package list with Pkg.update() before installing it, if you want to
try it out.

Cheers,
Mosè

Reply via email to