I believe delta BIC is calculated as:
delta BIC = BIC of the candidate model - lowest BIC value among all models
so that models with large delta BIC will not be favored.
So I'm surprised when I see that the vertical axis is showing the significant signal is having larger delta BIC value than other insignificant peaks in periodogram:
dbic1.pdf
Shouldn't the vertical axis be flipped?
Also, for some reasons one of my periodograms is returning negative delta BIC values and not showing FAP level:
dbic3.pdf
I believe delta BIC is calculated as: delta BIC = BIC of the candidate model - lowest BIC value among all models so that models with large delta BIC will not be favored. So I'm surprised when I see that the vertical axis is showing the significant signal is having larger delta BIC value than other insignificant peaks in periodogram: dbic1.pdf Shouldn't the vertical axis be flipped?
Also, for some reasons one of my periodograms is returning negative delta BIC values and not showing FAP level: dbic3.pdf