Sankhya: The Indian Journal of Statistics

2006, Volume 68, Pt. 2, 283--306

An Efficient and Fast Algorithm for Estimating the Parameters of Sinusoidal Signals

Swagata Nandi, Indian Statistical Institute, New Delhi, India
Debasis Kundu, Indian Institute of Technology, Kanpur, India

SUMMARY. A computationally efficient algorithm is proposed for estimating the parameters of sinusoidal signals in presence of stationary errors. The proposed estimators are consistent, and they are asymptotically equivalent to the least squares estimators. Monte Carlo simulations are performed to compare the proposed one with the other existing comparable methods. It is observed that the proposed estimator works quite well in terms of biases and mean squared errors. The main advantage of the proposed method of estimation is that the estimators can be obtained using only fixed number of iterations. Some real data sets have been analyzed for illustration purposes.

AMS (2000) subject classification. 62F10, 60K40.

Key words and phrases. Sinusoidal frequency model, least squares estimators, approximate least squares estimators, consistent estimators, asymptotic distributions, iterations, convergence.

Full paper (PDF)