Would you please recommend correct method of LPV color indexes calculation from scattered data? For example, light curves and phase plots of well known RVa EP Lyr have clear shape differences and shifts of extremums. How to calculate (B-V)=f(phase) or “delay” phase(V)-phase(B-V)?
I used absolutely awful polynomial interpolation and suspect it is wrong way.
Regretfully, I did not find suitable functions in Peranso. May VStar help? Or maybe there is ready application for such kind of analysis?
I tried to find indexes with absolutely stupid manual method finally, but it
is obviously wrong way. I’m confident there is normal solution of the “problem”.
(I use LibreOffice freeware, which returns polynomial coeff-ts in short format, so calculation of averaged BgVI values is impossble!)
Hmmm I found in VStar B-V plugin and tested it on EP Lyr AAVSO BV data.
I did not fully unterstand algorytm, but plugin works and gives B-V phase plot with smoothed polynomial approximation.
Do you want to get a simple average or median over an entire cycle? I presume not but if you do then you would take you phase data B-V say and take the median of all the B-V points. Assumes you have sampled the entire light curve well, without any major gaps in the data, especially near extrema. Your B-V plot is probably OK.
If you want to get the B-V, for example, more often during a cycle, then you could evaluate your smoothed model curves at the same phase and do a simple B-V difference at that phase, stepping from 0 to 1 every 0.1 or 0.05 phase or …
Excel, maybe other software, have add-ons you can get to do things like regressions and such with more precision and accuracy.
Thank you, Jim! Indeed, smooth fitting of whole cycle is not trivial task, and simpier to average data in the narrow ranges - if there are enough points.
Agree, Excel or MathCad will solve this problem more better than free Libre, but the best way would be VStar or Peranso
