[vsnet-chat 869] Re: Magnitude Estimates for Current SNe

[Taichi Kato <tkato>]

Re: [vsnet-chat 858] Re: Magnitude Estimates for Current SNe

  Upon question from Peter Nelson:

> How does kojinsa.exe operate?  Does it assume there is a correct value of
> magnitude or does it somehow average multiple observers with weightings?
> Where can it be downloaded?
> 
> Kind regards
> 
> Peter Nelson

   kojinsa.exe ("kojinsa" in Japanese means "personal difference") is a
DOS program, specially adapted for our domestic PC architecture.  So it will
not run on DOS/V machines.  As an alternative, I explain the algorithm.

   Assume a person_i has a personal trend (deviation) t_i from the ensemble
average of visual observations.  Then person_i's observation m_j may be
corrected to m_j - t_i.  The problem is how to determine the best estimate
of t_i.  I adopt the strategy so that the resultant light curve using
(m_j - t_i) for all observers becomes the smoothest.  The degree of smoothness
can be, as I formulated, estimated summing up square differences between
the adjacent (in time) observations (the similar strategy is also adopted
in period search using periodogram).  Though mathematically not tightly
linked, the smoothness above may represent something like a sum of local
scatter.  Minimizing this sum will provide a best estimate set of t_i's
on this formulation.

Regards,
Taichi Kato

VSNET Home Page

Return to Daisaku Nogami

---

vsnet-adm@kusastro.kyoto-u.ac.jp