Abstract
This paper discusses the theory and assumptions involved in the use of periodograms for seeking periodicities in astronomical time series. It is shown that previously derived formula for ‘false alarm’ probabilities severely underestimates the potential statistical errors, and that the most often used statistical hypothesis is inappropriate for testing for the presence of specific periodicities. An attempt was made to quantify the concept of ‘independent frequencies’, but the term is found to be misleading. It is suggested that the term needs reformulation.
Author
Koen, Chris
Journal
Astrophysical Journal
Paper Publication Date
January 1990
Paper Type
Astrostatistics