B‐spline approximation for the baseline hazard function

James Angelos, Carl M.‐S Lee, Karan P. Singh

Research output: Contribution to journalArticlepeer-review

1 Scopus citations


Etezadi‐Amoli and Ciampi (Biometrics 1987) introduced a method to approximate the baseline hazard and to estimate the regression coefficients for covariates simultaneously for the extended hazard regression (EHR) model using a quadratic spline approach. In this paper, an estimate of the baseline hazard function by B‐spline approximation using a minimax criterion is proposed. The nonlinear problem is approximated by a linear programming problem with only linear constraints. The nice features of this approach are: (i) the minimax criterion provides a robust approximation to the hazard function, and (ii) the linearized problem is numerically simple and fast.

Original languageEnglish
Pages (from-to)323-339
Number of pages17
Issue number3
StatePublished - 1991
Externally publishedYes


  • Extended hazard regression
  • censored survival data
  • linear programming
  • minimax criterion


Dive into the research topics of 'B‐spline approximation for the baseline hazard function'. Together they form a unique fingerprint.

Cite this