cumhaz               package:LogicReg               R Documentation

_C_u_m_u_l_a_t_i_v_e _h_a_z_a_r_d _t_r_a_n_s_f_o_r_m_a_t_i_o_n

_D_e_s_c_r_i_p_t_i_o_n:

     Transforms survival times using the cumulative hazard function.

_U_s_a_g_e:

     cumhaz(y, d)

_A_r_g_u_m_e_n_t_s:

       y: vector of nonnegative survival times

       d: vector of censoring indicators, should be the same length as
          'y'. If 'd' is missing the data is assumed to be uncensored.

_V_a_l_u_e:

     A vector of transformed survival times.

_N_o_t_e:

     The primary use of doing a cumulative hazard transformation is
     that after such a transformation, exponential survival models
     yield results that are often very much comparable to proportional
     hazards models. In our implementation of Logic Regression,
     however, exponential survival models run much faster than
     proportional hazards models when there are no continuous separate
     covariates.

_A_u_t_h_o_r(_s):

     Ingo Ruczinski ingo@jhu.edu and Charles Kooperberg clk@fhcrc.org.

_R_e_f_e_r_e_n_c_e_s:

     Ruczinski I, Kooperberg C, LeBlanc ML (2003).  Logic Regression,
     _Journal of Computational and Graphical Statistics_, *12*,
     475-511.

_S_e_e _A_l_s_o:

     'logreg'

_E_x_a_m_p_l_e_s:

     data(logreg.testdat)
     #
     # this is not survival data, but it shows the functionality
     yy <- cumhaz(exp(logreg.testdat[,1]), logreg.testdat[, 2])
     # then we would use
     # logreg(resp=yy, cens=logreg.testdat[,2], type=5, ...
     # insted of
     # logreg(resp=logreg.testdat[,1], cens=logreg.testdat[,2], type=4, ...

