Wavelet detection of change points in hazard rate models with censored dependent data

Abstract

The detection of change points in hazard rate has been studied a lot for independently and identically distributed survival data. However, in some domains, the survival times may be dependent. This paper considers the detection and estimation of change points in hazard rate for censored dependent data. We construct a nonparametric test statistic based on the wavelet method for change point detection. We also utilise the test statistic to design estimators for the number, the locations, and the jump sizes of the change points in hazard rate. The corresponding asymptotic properties are derived. Some simulation studies are conducted to assess the finite sample performances of the proposed method.

Keywords:

• change points
• hazard rate
• random censoring
• α-mixing
• wavelet coefficient
• kernel estimation

AMS Subject Classification :

• 62N02
• 62G05

Acknowledgements

The authors thank an anonymous reviewer and the associate editor whose insightful comments have resulted in a much improved paper. This research was supported by the National Natural Science Foundation of China (10971033).