On a Unification of Bias Reduction and Numerical Approximation.
SOUTHERN METHODIST UNIV DALLAS TEX DEPT OF STATISTICS
Pagination or Media Count:
In this paper it is shown that the problem of numerical approximation and bias reduction are basically the same problem and that many of the classical numerical methods are equivalent to the so-called jackknife method. In particular it is shown that Simpsons rule, Romberg integration, Newton-Cotes methods, Lagrange interpolation, the epsilon-algorithm, G-transforms, and others are simply special cases of the generalized jackknife. These observations are then used to obtain a new consistent estimator for the spectral density function. Author
- Statistics and Probability