Nonparametric estimation of hazard functions by wavelet methods
Samuel Shangwu Wu, Martin T. Wells · Journal of nonparametric statistics · 2003
This paper studies hazard rate estimation by non-linear wavelet methods in the setting of the counting process intensity model. An asymptotic formula for the mean integrated squared error (MISE) is provided. In addition, we show that the same MISE convergence result holds for hazard rates which are only piecewise smooth, in contrast to the analogous situation for kernel smoothing. Local properties, including consistency and asymptotic normality, are also established.