Bandwidth selection in marker dependent kernel hazard estimation

Practical estimation procedures for the local linear estimation of an unrestricted failure rate when more information is available than just time are developed. This extra information could be a covariate and this covariate could be a time series. Time dependent covariates are sometimes called markers, and failure rates are sometimes called hazards, intensities or mortalities. It is shown through simulations and a practical example that the fully local linear estimation procedure exhibits an excellent practical performance. Two different bandwidth selection procedures are developed. One is an adaptation of classical cross-validation, and the other one is indirect cross-validation. The simulation study concludes that classical cross-validation works well on continuous data while indirect cross-validation performs only marginally better. However, cross-validation breaks down in the practical data application to old-age mortality. Indirect cross-validation is thus shown to be superior when selecting a fully feasible estimation method for marker dependent hazard estimation.

[1]  Nader Ebrahimi,et al.  A semi-parametric generalization of the Cox proportional hazards regression model: Inference and applications , 2011, Comput. Stat. Data Anal..

[2]  Jens Perch Nielsen,et al.  Do-Validation for Kernel Density Estimation , 2011 .

[3]  Hyemi Choi,et al.  A local linear estimation of conditional hazard function in censored data , 2010 .

[4]  Laura Spierdijk,et al.  Nonparametric conditional hazard rate estimation: A local linear approach , 2008, Comput. Stat. Data Anal..

[5]  Haifen Li,et al.  A new semiparametric estimation method for accelerated hazards mixture cure model , 2013, Comput. Stat. Data Anal..

[6]  E. Mammen,et al.  A General Approach to the Predictability Issue in Survival Analysis with Applications , 2007 .

[7]  O. Linton,et al.  Kernel estimation in a nonparametric marker dependent hazard model , 1995 .

[8]  One Sided Cross Validation for Density Estimation with an Application to Operational Risk , 2006 .

[9]  Jens Perch Nielsen,et al.  Boundary and Bias Correction in Kernel Hazard Estimation , 2001 .

[10]  J. Nielsen,et al.  Multivariate Density Estimation Using Dimension Reducing Information and Tail Flattening Transformations for Truncated or Censored Data , 2010 .

[11]  J. Nielsen.  Variable bandwidth kernel hazard estimators , 2003 .

[12]  Niels Keiding,et al.  Statistical Models Based on Counting Processes , 1993 .

[13]  D. Cox.  Regression Models and Life-Tables , 1972 .

[14]  David R. Cox,et al.  Regression models and life tables (with discussion , 1972 .

[15]  Shean-Tsong Chiu,et al.  Bandwidth selection for kernel density estimation , 1991 .

[16]  M. C. Jones,et al.  Simple boundary correction for kernel density estimation , 1993 .

[17]  María Dolores Martínez Miranda,et al.  Smoothing survival densities in practice , 2013, Comput. Stat. Data Anal..

[18]  Some aspects of hadamard differentiability on regression L-estimators , 1994 .

[19]  D. Bagkavos.  Local linear hazard rate estimation and bandwidth selection , 2011 .

[20]  Jens Perch Nielsen,et al.  Marker dependent kernel hazard estimation from local linear estimation , 1998 .

[21]  R. Prentice,et al.  Commentary on Andersen and Gill's "Cox's Regression Model for Counting Processes: A Large Sample Study" , 1982 .

[22]  T. Martinussen,et al.  The additive hazards model with high-dimensional regressors , 2009, Lifetime data analysis.

[23]  O. Aalen.  Nonparametric Inference for a Family of Counting Processes , 1978 .

[24]  J. Nielsen,et al.  A nonparametric visual test of mixed hazard models , 2013 .

[25]  Simon J. Sheather,et al.  Indirect Cross-Validation for Density Estimation , 2008, 0812.0051.

[26]  J. Nielsen,et al.  A Comparative Study of Parametric and Nonparametric Estimators of Old-Age Mortality in Sweden , 2004 .

[27]  Haifen Li,et al.  Induced smoothing for the semiparametric accelerated hazards model , 2012, Comput. Stat. Data Anal..

[28]  R Core Team,et al.  R: A language and environment for statistical computing. , 2014 .

[29]  Dimitrios Bagkavos,et al.  Local Polynomial Fitting in Failure Rate Estimation , 2008, IEEE Transactions on Reliability.