## The Annals of Statistics

### Local likelihood and local partial likelihood in hazard regression

#### Abstract

In survival analysis, the relationship between a survival time and a covariate is conveniently modeled with the proportional hazards regression model. This model usually assumes that the covariate has a log-linear effect on the hazard function. In this paper we consider the proportional hazards regression model with a nonparametric risk effect. We discuss estimation of the risk function and its derivatives in two cases: when the baseline hazard function is parametrized and when it is not parametrized. In the case of a parametric baseline hazard function, inference is based on a local version of the likelihood function, while in the case of a nonparametric baseline hazard, we use a local version of the partial likelihood. This results in maximum local likelihood estimators and maximum local partial likelihood estimators, respectively. We establish the asymptotic normality of the estimators. It turns out that both methods have the same asymptotic bias and variance in a common situation, even though the local likelihood method uses information about the baseline hazard function.

#### Article information

Source
Ann. Statist., Volume 25, Number 4 (1997), 1661-1690.

Dates
First available in Project Euclid: 9 September 2002

https://projecteuclid.org/euclid.aos/1031594736

Digital Object Identifier
doi:10.1214/aos/1031594736

Mathematical Reviews number (MathSciNet)
MR1463569

Zentralblatt MATH identifier
0890.62023

Subjects
Primary: 62G05: Estimation
Secondary: 62E20: Asymptotic distribution theory 60G44: Martingales with continuous parameter

#### Citation

Fan, Jianqing; Gijbels, Irène; King, Martin. Local likelihood and local partial likelihood in hazard regression. Ann. Statist. 25 (1997), no. 4, 1661--1690. doi:10.1214/aos/1031594736. https://projecteuclid.org/euclid.aos/1031594736

#### References

• AITKIN, M. and CLAy TON, D. G. 1980. The fitting of exponential, Weibull and extreme value distributions to complex censored survival data using GLIM. J. Roy. Statist. Soc. Ser. C 29 156 163. Z.
• ANDERSEN, P. K., BORGAN, Ø., GILL, R. D. and KEIDING, N. 1993. Statistical Models Based on Counting Processes. Springer, New York. Z.
• ANDERSEN, P. K. and GILL, R. D. 1982. Cox's regression model for counting processes: a large sample study. Ann. Statist. 10 1100 1120. Z.
• BRESLOW, N. E. 1972. Comment on Regression and life tables,'' by D. R. Cox. J. Roy. Statist. Soc. Ser. B 34 216 217. Z.
• BRESLOW, N. E. 1974. Covariance analysis of censored survival data. Biometrics 30 89 99. Z. Z.
• COX, D. R. 1972. Regression models and life-tables with discussion. J. Roy. Statist. Soc. Ser. B 4 187 220. Z.
• COX, D. R. 1975. Partial likelihood. Biometrika 62 269 276. Z.
• FAN, J. and GIJBELS, I. 1996. Local Poly nomial Modelling and Its Applications. Chapman and Hall, London. Z.
• FAN, J., HARDLE, W. and MAMMEN, E. 1995. Direct estimation of additive and linear compo¨nents for high dimensional data. Inst. Statist. Mimeo Series 2339. Univ. North Carolina, Chapel Hill. Z.
• FLEMING, T. R. and HARRINGTON, D. P. 1991. Counting Processes and Survival Analy sis. Wiley, New York. Z.
• GENTLEMAN, R. and CROWLEY, J. 1991. Local full likelihood estimation for the proportional hazards model. Biometrics 47 1283 1296. Z.
• HASTIE, T. and TIBSHIRANI, R. 1990a. Generalized Additive Models. Chapman and Hall, London. Z.
• HASTIE, T. and TIBSHIRANI, R. 1990b. Exploring the nature of covariate effects in the proportional hazards model. Biometrics 46 1005 1016. Z.
• HJORT, N. L. 1996. Dy namic likelihood hazard rate estimation. Biometrika. To appear. Z.
• KOOPERBERG, C., STONE, C. J. and TRUONG, Y. 1995a. Hazard regression. J. Amer. Statist. Assoc. 90 78 94. Z.
• KOOPERBERG, C., STONE, C. J. and TRUONG, Y. 1995b. The L rate of convergence for hazard 2 regression. Scand. J. Statist. 22 143 157. Z.
• LEHMANN, E. L. 1983. Theory of Point Estimation. Wadsworth & Brooks Cole, Pacific Grove, CA. Z.
• LI, G. and DOSS, H. 1995. An approach to nonparametric regression for life history data using local linear fitting. Ann. Statist. 23 787 823. Z.
• LINTON, O. and NIELSEN, J. P. 1995. A kernel method of estimating structured nonparametric regression based on marginal integration. Biometrika 82 93 100. Z.
• MARRON, J. S. and NOLAN, D. 1988. Canonical kernels for density estimation. Statist. Probab. Lett. 7 195 199. Z.
• MARRON, J. S. and PADGETT, W. J. 1987. Asy mptotically optimal bandwidth selection from randomly right-censored samples. Ann. Statist. 15 1520 1535. Z.
• MULLER, H. G. and WANG, J. L. 1990. Analy zing changes in hazard functions: an alternative to ¨ change-point models. Biometrika 77 610 625. Z.
• MULLER, H. G. and WANG, J. L. 1994. Hazard rate estimation under random censoring with ¨ varying kernels and bandwidths. Biometrics 50 61 76. Z. O'SULLIVAN, F. 1988. Nonparametric estimation of relative risk using splines and crossvalidation. SIAM J. Sci. Statist. Comput. 9 531 542. Z.
• POLLARD, D. 1984. Convergence of Stochastic Processes. Springer, New York. Z.
• STONE, C. J. 1994. The use of poly nomial splines and their tensor products in multivariate Z. function estimation with discussion. Ann. Statist. 22 118 184.
• STUTE, W. and WANG, J. L. 1993. A strong law under random censorship. Ann. Statist. 21 1591 1607. Z.
• TIBSHIRANI, R. and HASTIE, T. 1987. Local likelihood estimation. J. Amer. Statist. Assoc. 82 559 567. Z.
• WONG, W. H. 1986. Theory of partial likelihood. Ann. Statist. 14 88 123.
• CHAPEL HILL, NORTH CAROLINA 27599-3260