| lrm.fit {Design} | R Documentation |
Fits a binary or ordinal logistic model for a given design matrix and response vector with no missing values in either. Ordinary or penalized maximum likelihood estimation is used.
lrm.fit(x, y, offset, initial, est, maxit=12, eps=.025,
tol=1E-7, trace=FALSE, penalty.matrix, weights, normwt)
x |
design matrix with no column for an intercept |
y |
response vector, numeric, categorical, or character |
offset |
optional numeric vector containing an offset on the logit scale |
initial |
vector of initial parameter estimates, beginning with the intercept |
est |
indexes of x to fit in the model (default is all columns of x).
Specifying est=c(1,2,5) causes columns 1,2, and 5 to have
parameters estimated. The score vector u and covariance matrix var
can be used to obtain score statistics for other columns
|
maxit |
maximum no. iterations (default=12). Specifying maxit=1
causes logist to compute statistics at initial estimates.
|
eps |
difference in -2 log likelihood for declaring convergence.
Default is .025.
|
tol |
Singularity criterion. Default is 1E-7 |
trace |
set to TRUE to print -2 log likelihood, step-halving
fraction, and rank of variance matrix at each iteration
|
penalty.matrix |
a self-contained ready-to-use penalty matrix - see lrm
|
weights |
a vector (same length as y) of possibly fractional case weights
|
normwt |
set to code{TRUE} to scale weights so they sum to the length of
y; useful for sample surveys as opposed to the default of
frequency weighting
|
a list with the following components:
call |
calling expression |
freq |
table of frequencies for y in order of increasing y
|
stats |
vector with the following elements: number of observations used in the
fit, maximum absolute value of first
derivative of log likelihood, model likelihood ratio chi-square, d.f.,
P-value,
c index (area under ROC curve), Somers' D_{xy},
Goodman-Kruskal gamma, and Kendall's tau-a
rank correlations
between predicted probabilities and observed response, the
Nagelkerke R^2 index, and the Brier probability score with
respect to computing the probability that y > lowest level.
Probabilities are rounded to the nearest 0.002
in the computations or rank correlation indexes.
When penalty.matrix is present, the chi-square,
d.f., and P-value are not corrected for the effective d.f.
|
fail |
set to TRUE if convergence failed (and maxiter>1)
|
coefficients |
estimated parameters |
var |
estimated variance-covariance matrix (inverse of information matrix).
Note that in the case of penalized estimation, var is not the
improved sandwich-type estimator (which lrm does compute).
|
u |
vector of first derivatives of log-likelihood |
deviance |
-2 log likelihoods. When an offset variable is present, three deviances are computed: for intercept(s) only, for intercepts+offset, and for intercepts+offset+predictors. When there is no offset variable, the vector contains deviances for the intercept(s)-only model and the model with intercept(s) and predictors. |
est |
vector of column numbers of X fitted (intercepts are not counted)
|
non.slopes |
number of intercepts in model |
penalty.matrix |
see above |
Frank Harrell
Department of Biostatistics, Vanderbilt University
f.harrell@vanderbilt.edu
lrm, glm, matinv, solvet, cr.setup
#Fit an additive logistic model containing numeric predictors age, #blood.pressure, and sex, assumed to be already properly coded and #transformed # # fit <- lrm.fit(cbind(age,blood.pressure,sex), death)