The detection of hidden periodicities: A comparison of alternative methods

"Fixed frequency effect models" represent a powerful tool for analyzing time series exhibiting strong periodicities. However, in spite of their appeal to the practitioner, their use has been constrained by ignorance about their statistical properties. This paper attempts to oer a comparison among alternative methods via extensive simulation studies. The methods are compared across several performance characteristics most notably bias, variance power and RMSE (root mean square error). By way of illustration, two empirical examples are also included.