Interface CoxphResult

Coefficients (NaN for R's NA: an aliased column or one whose information fell below the tolerance)

Number of iterations (iterMax + 1 when they ran out)

Linear predictors (centred by the means), one per observation used, in input order

Log partial likelihood at the starting values and at the end

Likelihood ratio test

Lower confidence limits of exp(coef) (from robustSe when usedRobust)

The means used to centre the covariates (0 for the columns not centred)

Number of observations used

The model-based variance (R's naive.var); only when usedRobust.

Number of events

Two-sided p-values Pr(>|z|)

Alias of p: the same array.

The robust score test (R's robscore: rscore, the number of non-NaN coefficients); only when usedRobust.

Robust standard errors, the square roots of the diagonal of var (R's robust se); only when usedRobust.

The robust score test statistic at init (R's rscore); only when usedRobust.

R^2 (1 - exp(-logtest / n)) and its maximum (1 - exp(2 loglik[0] / n))

The score test statistic at the starting values

Score (log-rank) test

Model-based (naive) standard errors of the coefficients, also when usedRobust (R's column se(coef); the robust ones are robustSe)

Alias of se: the same array.

Upper confidence limits of exp(coef) (from robustSe when usedRobust)

Whether the robust variance was used (R's used.robust). False also when it was asked for but there is nothing to compute it from (no events, or every coefficient NaN), as R.

Variance matrix of the coefficients (array of rows): the robust variance when usedRobust (R's var), else the model-based one

The Wald statistic of the non-NA coefficients (against init when given)

Wald test (with the robust variance when usedRobust)

R's warnings, each once, in order of first appearance.

Wald z statistics (coefficient / robustSe when usedRobust, else / se)

Alias of z: the same array.