help logistic postestimation dialogs: predict estat
lroc lsens
also see: logistic
-------------------------------------------------------------------------------
Title
[R] logistic postestimation -- Postestimation tools for logistic
Description
The following postestimation commands are of special interest after
logistic:
command description
-------------------------------------------------------------------------
estat classification reports various summary statistics, including the
classification table
estat gof Pearson or Hosmer-Lemeshow goodness-of-fit test
lroc graphs the ROC curve and calculates the area under
the curve
lsens graphs sensitivity and specificity versus
probability cutoff
-------------------------------------------------------------------------
These commands are not appropriate after the svy prefix.
The following standard postestimation commands are also available:
command description
-------------------------------------------------------------------------
estat AIC, BIC, VCE, and estimation sample summary
estat (svy) postestimation statistics for survey data
estimates cataloging estimation results
lincom point estimates, standard errors, testing, and
inference for linear combinations of coefficients
linktest link test for model specification
(1) lrtest likelihood-ratio test
margins marginal means, predictive margins, marginal effects,
and average marginal effects
nlcom point estimates, standard errors, testing, and
inference for nonlinear combinations of coefficients
predict predictions, residuals, influence statistics, and other
diagnostic measures
predictnl point estimates, standard errors, testing, and
inference for generalized predictions
suest seemingly unrelated estimation
test Wald tests of simple and composite linear hypotheses
testnl Wald tests of nonlinear hypotheses
-------------------------------------------------------------------------
(1) lrtest is not appropriate with svy estimation results.
Special-interest postestimation commands
estat classification reports various summary statistics, including the
classification table.
estat gof reports the Pearson goodness-of-fit test or the Hosmer-Lemeshow
goodness-of-fit test.
lroc graphs the ROC curve and calculates the area under the curve.
lsens graphs sensitivity and specificity versus probability cutoff and
optionally creates new variables containing these data.
estat classification, estat gof, lroc, and lsens produce statistics and
graphs either for the estimation sample or for any set of observations.
However, they always use the estimation sample by default. When weights,
if, or in is used with logistic, it is not necessary to repeat the
qualifier with these commands when you want statistics computed for the
estimation sample. Specify if, in, or the all option only when you want
statistics computed for a set of observations other than the estimation
sample. Specify weights (only fweights are allowed with these commands)
only when you want to use a different set of weights.
By default, estat classification, estat gof, lroc, and lsens use the last
model fit by logistic. You may also directly specify the model to the
lroc and lsens commands by inputting a vector of coefficients with the
beta() option and passing the name of the dependent variable depvar.
estat classification and estat gof require that the current estimation
results be from logistic, logit, or probit. lroc and lsens commands may
also be used after logit or probit.
Syntax for predict
predict [type] newvar [if] [in] [, statistic nooffset rules asif]
statistic description
-------------------------------------------------------------------------
Main
pr probability of a positive outcome; the default
xb linear prediction
stdp standard error of the prediction
* dbeta Pregibon Delta-Beta influence statistic
* deviance deviance residual
* dx2 Hosmer and Lemeshow Delta chi-squared influence statistic
* ddeviance Hosmer and Lemeshow Delta-D influence statistic
* hat Pregibon leverage
* number sequential number of the covariate pattern
* residuals Pearson residuals; adjusted for number sharing covariate
pattern
* rstandard standardized Pearson residuals; adjusted for number
sharing covariate pattern
score first derivative of the log likelihood with respect to
x-Beta
-------------------------------------------------------------------------
Unstarred statistics are available both in and out of sample; type
predict ... if e(sample) ... if wanted only for the estimation sample.
Starred statistics are calculated only for the estimation sample, even
when if e(sample) is not specified.
pr, xb, stdp, and score are the only options allowed with svy estimation
results.
Menu
Statistics > Postestimation > Predictions, residuals, etc.
Options for predict
+------+
----+ Main +-------------------------------------------------------------
pr, the default, calculates the probability of a positive outcome.
xb calculates the linear prediction.
stdp calculates the standard error of the linear prediction.
dbeta calculates the Pregibon Delta-Beta influence statistic, a
standardized measure of the difference in the coefficient vector that
is due to deletion of the observation along with all others that
share the same covariate pattern. In Hosmer and Lemeshow jargon,
this statistic is M-asymptotic; that is, it is adjusted for the
number of observations that share the same covariate pattern.
deviance calculates the deviance residual.
dx2 calculates the Hosmer and Lemeshow Delta chi-squared influence
statistic, reflecting the decrease in the Pearson chi-squared because
of deletion of the observation and all others that share the same
covariate pattern.
ddeviance calculates the Hosmer and Lemeshow Delta-D influence statistic,
which is the change in the deviance residual that is due to deletion
of the observation and all others that share the same covariate
pattern.
hat calculates the Pregibon leverage or the diagonal elements of the hat
matrix adjusted for the number of observations that share the same
covariate pattern.
number numbers the covariate patterns -- observations with the same
covariate pattern have the same number. Observations not used in
estimation have number set to missing. The first covariate pattern
is numbered 1, the second 2, and so on.
residuals calculates the Pearson residual as given by Hosmer and Lemeshow
and adjusted for the number of observations that share the same
covariate pattern.
rstandard calculates the standardized Pearson residual as given by Hosmer
and Lemeshow and adjusted for the number of observations that share
the same covariate pattern.
score calculates the equation-level score, the first derivative of the
log likelihood with respect to the linear prediction.
+---------+
----+ Options +----------------------------------------------------------
nooffset is relevant only if you specified offset(varname) for logistic.
It modifies the calculations made by predict so that they ignore the
offset variable; the linear prediction is treated as xb rather than
xb + offset.
rules requests that Stata use any rules that were used to identify the
model when making the prediction. By default, Stata calculates
missing for excluded observations. See [R] logit for an example.
asif requests that Stata ignore the rules and the exclusion criteria and
calculate predictions for all observations possible by using the
estimated parameter from the model. See [R] logit for an example.
Syntax for estat classification
estat classification [if] [in] [weight] [, class_options]
class_options description
-------------------------------------------------------------------------
Main
all display summary statistics for all observations in the
data
cutoff(#) positive outcome threshold; default is cutoff(0.5)
-------------------------------------------------------------------------
fweights are allowed; see weight.
Menu
Statistics > Postestimation > Reports and statistics
Options for estat classification
+------+
----+ Main +-------------------------------------------------------------
all requests that the statistic be computed for all observations in the
data, ignoring any if or in restrictions specified by logistic.
cutoff(#) specifies the value for determining whether an observation has
a predicted positive outcome. An observation is classified as
positive if its predicted probability is > #. The default is 0.5.
Syntax for estat gof
estat gof [if] [in] [weight] [, gof_options]
gof_options description
-------------------------------------------------------------------------
Main
group(#) perform Hosmer-Lemeshow goodness-of-fit test using #
quantiles
all execute test for all observations in the data
outsample adjust degrees of freedom for samples outside estimation
sample
table display table of groups used for test
-------------------------------------------------------------------------
fweights are allowed; see weight.
Menu
Statistics > Postestimation > Reports and statistics
Options for estat gof
+------+
----+ Main +-------------------------------------------------------------
group(#) specifies the number of quantiles to be used to group the data
for the Hosmer-Lemeshow goodness-of-fit test. group(10) is typically
specified. If this option is not given, the Pearson goodness-of-fit
test is computed using the covariate patterns in the data as groups.
all requests that the statistic be computed for all observations in the
data, ignoring any if or in restrictions specified with logistic.
outsample adjusts the degrees of freedom for the Pearson and
Hosmer-Lemeshow goodness-of-fit tests for samples outside the
estimation sample.
table displays a table of the groups used for the Hosmer-Lemeshow or
Pearson goodness-of-fit test with predicted probabilities, observed
and expected counts for both outcomes, and totals for each group.
Syntax for lroc
lroc [depvar] [if] [in] [weight] [, lroc_options]
lroc_options description
-------------------------------------------------------------------------
Main
all compute area under ROC curve and graph curve for
all observations
nograph suppress graph
Advanced
beta(matname) row vector containing coefficients for a
logistic model
Plot
cline_options change the look of the line
marker_options change look of markers (color, size, etc.)
marker_label_options add marker labels; change look or position
Reference line
rlopts(cline_options) affect rendition of the reference line
Add plots
addplot(plot) add other plots to the generated graph
Y axis, X axis, Titles, Legend, Overall
twoway_options any options other than by() documented in
[G] twoway_options
-------------------------------------------------------------------------
fweights are allowed; see weight.
Menu
Statistics > Binary outcomes > Postestimation > ROC curve after
logistic/logit/probit/ivprobit
Options for lroc
+------+
----+ Main +-------------------------------------------------------------
all requests that the statistic be computed for all observations in the
data, ignoring any if or in restrictions specified by logistic.
nograph suppresses graphical output.
+----------+
----+ Advanced +---------------------------------------------------------
beta(matname) specifies a row vector containing coefficients for a
logistic model. The columns of the row vector must be labeled with
the corresponding names of the independent variables in the data.
The dependent variable depvar must be specified immediately after the
command name.
+------+
----+ Plot +-------------------------------------------------------------
cline_options, marker_options, and marker_label_options affect the
rendition of the ROC curve -- the plotted points connected by lines.
These options affect the size and color of markers, whether and how
the markers are labeled, and whether and how the points are
connected; see [G] cline_options, [G] marker_options, and [G]
marker_label_options.
+----------------+
----+ Reference line +---------------------------------------------------
rlopts(cline_options) affects the rendition of the reference line; see
[G] cline_options.
+-----------+
----+ Add plots +--------------------------------------------------------
addplot(plot) provides a way to add other plots to the generated graph;
see [G] addplot_option.
+-----------------------------------------+
----+ Y axis, X axis, Titles, Legend, Overall +--------------------------
twoway_options are any of the options documented in [G] twoway_options,
excluding by(). These include options for titling the graph (see [G]
title_options) and for saving the graph to disk (see [G]
saving_option).
Syntax for lsens
lsens [depvar] [if] [in] [weight] [, lsens_options]
lsens_options description
-------------------------------------------------------------------------
Main
all graph all observations in the data
genprob(varname) create variable containing probability cutoff
gensens(varname) create variable containing sensitivity
genspec(varname) create variable containing specificity
replace overwrite existing variables
nograph suppress the graph
Advanced
beta(matname) row vector containing coefficients for the model
Plot
connect_options affect rendition of the plotted points connected by
lines
Add plots
addplot(plot) add other plots to the generated graph
Y axis, X axis, Titles, Legend, Overall
twoway_options any options other than by() documented in
[G] twoway_options
-------------------------------------------------------------------------
fweights are allowed; see weight.
Menu
Statistics > Binary outcomes > Postestimation > Sensitivity/specificity
plot
Options for lsens
+------+
----+ Main +-------------------------------------------------------------
all requests that the statistic be computed for all observations in the
data, ignoring any if or in restrictions specified with logistic.
genprob(varname), gensens(varname), and genspec(varname) specify the
names of new variables created to contain, respectively, the
probability cutoffs and the corresponding sensitivity and
specificity.
replace requests that existing variables specified for genprob(),
gensens(), or genspec() be overwritten.
nograph suppresses graphical output.
+----------+
----+ Advanced +---------------------------------------------------------
beta(matname) specifies a row vector containing coefficients for a
logistic model. The columns of the row vector must be labeled with
the corresponding names of the independent variables in the data.
The dependent variable depvar must be specified immediately after the
command name.
+------+
----+ Plot +-------------------------------------------------------------
connect_options affect the rendition of the plotted points connected by
lines; see connect_options in [G] graph twoway scatter.
+-----------+
----+ Add plots +--------------------------------------------------------
addplot(plot) provides a way to add other plots to the generated graph.
See [G] addplot_option.
+-----------------------------------------+
----+ Y axis, X axis, Titles, Legend, Overall +--------------------------
twoway_options are any of the options documented in [G] twoway_options,
excluding by(). These include options for titling the graph (see [G]
title_options) and for saving the graph to disk (see [G]
saving_option).
Examples
. webuse lbw
. logistic low age lwt i.race smoke ptl ht ui
. lroc
. lsens
. estat class
. estat gof
. predict phat if e(sample)
. predict r, resid
Saved results
estat classification saves the following in r():
Scalars
r(P_corr) percent correctly classified
r(P_p1) sensitivity
r(P_n0) specificity
r(P_p0) false-positive rate given true negative
r(P_n1) false-negative rate given true positive
r(P_1p) positive predictive value
r(P_0n) negative predictive value
r(P_0p) false-positive rate given classified positive
r(P_1n) false-negative rate given classified negative
estat gof saves the following in r():
Scalars
r(N) number of observations
r(m) number of covariate patterns or groups
r(df) degrees of freedom
r(chi2) chi-squared
lroc saves the following in r():
Scalars
r(N) number of observations
r(area) area under the ROC curve
lsens saves the following in r():
Scalars
r(N) number of observations
Also see
Manual: [R] logistic postestimation
Help: [R] logistic
