Bootstrapping bispectra: an application to testing for departure from Gaussianity of stationary signals

such an approach is local. We can choose the interpolation points nonuniformly by assigning more interpolation points around irregular locations. The disadvantage of ISS is that the generated curves are not polynomial splines. V. CONCLUSION In this correspondence, we propose the ISS approach for the initialization of a wavelet transform, which is very important for further applications. Moreover, the proposed method also provides an efficient evaluation of inner products of wavelets and their derivatives, which is essential in the wavelet–Galerkin method. We formulate a general procedure for the computation of exact derivative values of the interpolatory fundamental function at dyadic points. The error analysis shows that the method is exact for certain polynomials. The numerical experiments have shown the high accuracy and efficiency of the method. sampling series and the reconstruction of signals in terms of linear, quadrature and cubic splines, " SIAM J. Fast B-spline algorithms for continuous image representation and interpolation, " IEEE Trans. Fast implementation of scale-space filtering using interpolatory subdivision scheme, " IEEE Trans. Pattern Anal. Machine Intell., to be published. [10] S. Dubuc, " Interpolation through an iterative scheme, " J. Math. Anal. A 4-point interpolatory subdivision scheme for curve design, " Comput. , " Solving two point boundary value problems by interpolatory subdivision algorithms, " Int. Abstract— We propose a bootstrap version of a bispectrum-based test for departure from Gaussianity that achieves high power while maintaining the level of significance, even for small sample sizes. The proposed procedure can be also used to set confidence bands for a measure of the bicoherence of stationary random signals.