An efficient, noise tolerant, linear extrapolator
P.J. Thompson, MICHAEL T. MANRY · 2003
The extrapolation algorithms are commonly used to replace large intervals of bad or missing data, given knowledge of the original signal's autocovariance or cutoff frequencies. Present algorithms work well on idealized signals, but perform poorly when applied to noisy signals likely to be encountered in a real-world situation. A novel approach for extrapolation is developed which involves the solution of a Toeplitz set of linear equations. The advantages of the algorithm are that it is more efficient than its predecessors. It is not restricted to bandlimited signals and is linear.>