The Estimation of Spectral Densities
谱密度估计
SUMMARY A method of estimating spectral densities is developed which uses the data explicitly to help determine the best smoothing procedure. Two techniques are discussed. The first involves polynomial approximations, the second is an interpolation technique which is useful when a limited number of autocorrelations is available. The Gaussian and Laplacian weight functions play important roles in the first and second technique respectively.