CMStatistics 2022: Start Registration
View Submission - CMStatistics
Title: Frequency estimation of irregularly sampled time series with red noise Authors:  Efthymia Derezea - University of Kent (United Kingdom) [presenting]
Alfred Kume - University of Kent (United Kingdom)
Abstract: Irregularly sampled time series appear in many fields, such as Astronomy, Climatology, Economics etc. In many cases, it is of interest to estimate and identify possible periodic patterns. While many methods exist for the case of regularly sampled data, the problem is under-explored for time series with unequal time spacing. A simple harmonic model is considered with additional red noise for this type of data. The frequency estimate under this setting is proven to be consistent and asymptotically normal. This result is verified through an extensive simulation study which is reported along with an application to real data from Astronomy.