EVOLUTION-MANAGER
Edit File: augment.coxph.html
<!DOCTYPE html PUBLIC "-//W3C//DTD XHTML 1.0 Strict//EN" "http://www.w3.org/TR/xhtml1/DTD/xhtml1-strict.dtd"><html xmlns="http://www.w3.org/1999/xhtml"><head><title>R: Augment data with information from a(n) coxph object</title> <meta http-equiv="Content-Type" content="text/html; charset=utf-8" /> <link rel="stylesheet" type="text/css" href="R.css" /> </head><body> <table width="100%" summary="page for augment.coxph {broom}"><tr><td>augment.coxph {broom}</td><td style="text-align: right;">R Documentation</td></tr></table> <h2>Augment data with information from a(n) coxph object</h2> <h3>Description</h3> <p>Augment accepts a model object and a dataset and adds information about each observation in the dataset. Most commonly, this includes predicted values in the <code>.fitted</code> column, residuals in the <code>.resid</code> column, and standard errors for the fitted values in a <code>.se.fit</code> column. New columns always begin with a <code>.</code> prefix to avoid overwriting columns in the original dataset. </p> <p>Users may pass data to augment via either the <code>data</code> argument or the <code>newdata</code> argument. If the user passes data to the <code>data</code> argument, it <strong>must</strong> be exactly the data that was used to fit the model object. Pass datasets to <code>newdata</code> to augment data that was not used during model fitting. This still requires that all columns used to fit the model are present. </p> <p>Augment will often behave differently depending on whether <code>data</code> or <code>newdata</code> is given. This is because there is often information associated with training observations (such as influences or related) measures that is not meaningfully defined for new observations. </p> <p>For convenience, many augment methods provide default <code>data</code> arguments, so that <code>augment(fit)</code> will return the augmented training data. In these cases, augment tries to reconstruct the original data based on the model object with varying degrees of success. </p> <p>The augmented dataset is always returned as a <a href="../../tibble/html/tibble.html">tibble::tibble</a> with the <strong>same number of rows</strong> as the passed dataset. This means that the passed data must be coercible to a tibble. At this time, tibbles do not support matrix-columns. This means you should not specify a matrix of covariates in a model formula during the original model fitting process, and that <code><a href="../../splines/html/ns.html">splines::ns()</a></code>, <code><a href="../../stats/html/poly.html">stats::poly()</a></code> and <code><a href="../../survival/html/Surv.html">survival::Surv()</a></code> objects are not supported in input data. If you encounter errors, try explicitly passing a tibble, or fitting the original model on data in a tibble. </p> <p>We are in the process of defining behaviors for models fit with various <code>na.action</code> arguments, but make no guarantees about behavior when data is missing at this time. </p> <h3>Usage</h3> <pre> ## S3 method for class 'coxph' augment( x, data = NULL, newdata = NULL, type.predict = "lp", type.residuals = "martingale", ... ) </pre> <h3>Arguments</h3> <table summary="R argblock"> <tr valign="top"><td><code>x</code></td> <td> <p>A <code>coxph</code> object returned from <code><a href="../../survival/html/coxph.html">survival::coxph()</a></code>.</p> </td></tr> <tr valign="top"><td><code>data</code></td> <td> <p>A <a href="../../base/html/data.frame.html">base::data.frame</a> or <code><a href="../../tibble/html/tibble.html">tibble::tibble()</a></code> containing the original data that was used to produce the object <code>x</code>. Defaults to <code>stats::model.frame(x)</code> so that <code>augment(my_fit)</code> returns the augmented original data. <strong>Do not</strong> pass new data to the <code>data</code> argument. Augment will report information such as influence and cooks distance for data passed to the <code>data</code> argument. These measures are only defined for the original training data.</p> </td></tr> <tr valign="top"><td><code>newdata</code></td> <td> <p>A <code><a href="../../base/html/data.frame.html">base::data.frame()</a></code> or <code><a href="../../tibble/html/tibble.html">tibble::tibble()</a></code> containing all the original predictors used to create <code>x</code>. Defaults to <code>NULL</code>, indicating that nothing has been passed to <code>newdata</code>. If <code>newdata</code> is specified, the <code>data</code> argument will be ignored.</p> </td></tr> <tr valign="top"><td><code>type.predict</code></td> <td> <p>Character indicating type of prediction to use. Passed to the <code>type</code> argument of the <code><a href="../../stats/html/predict.html">stats::predict()</a></code> generic. Allowed arguments vary with model class, so be sure to read the <code>predict.my_class</code> documentation.</p> </td></tr> <tr valign="top"><td><code>type.residuals</code></td> <td> <p>Character indicating type of residuals to use. Passed to the <code>type</code> argument of <code><a href="../../stats/html/residuals.html">stats::residuals()</a></code> generic. Allowed arguments vary with model class, so be sure to read the <code>residuals.my_class</code> documentation.</p> </td></tr> <tr valign="top"><td><code>...</code></td> <td> <p>Additional arguments. Not used. Needed to match generic signature only. <strong>Cautionary note:</strong> Misspelled arguments will be absorbed in <code>...</code>, where they will be ignored. If the misspelled argument has a default value, the default value will be used. For example, if you pass <code>conf.lvel = 0.9</code>, all computation will proceed using <code>conf.level = 0.95</code>. Additionally, if you pass <code>newdata = my_tibble</code> to an <code><a href="reexports.html">augment()</a></code> method that does not accept a <code>newdata</code> argument, it will use the default value for the <code>data</code> argument.</p> </td></tr> </table> <h3>Details</h3> <p>When the modeling was performed with <code>na.action = "na.omit"</code> (as is the typical default), rows with NA in the initial data are omitted entirely from the augmented data frame. When the modeling was performed with <code>na.action = "na.exclude"</code>, one should provide the original data as a second argument, at which point the augmented data will contain those rows (typically with NAs in place of the new columns). If the original data is not provided to <code><a href="reexports.html">augment()</a></code> and <code>na.action = "na.exclude"</code>, a warning is raised and the incomplete rows are dropped. </p> <h3>Value</h3> <p>A <code><a href="../../tibble/html/tibble.html">tibble::tibble()</a></code> with columns: </p> <table summary="R valueblock"> <tr valign="top"><td><code>.fitted</code></td> <td> <p>Fitted or predicted value.</p> </td></tr> <tr valign="top"><td><code>.resid</code></td> <td> <p>The difference between observed and fitted values.</p> </td></tr> <tr valign="top"><td><code>.se.fit</code></td> <td> <p>Standard errors of fitted values.</p> </td></tr> </table> <h3>See Also</h3> <p><a href="../../stats/html/na.action.html">stats::na.action</a> </p> <p><code><a href="reexports.html">augment()</a></code>, <code><a href="../../survival/html/coxph.html">survival::coxph()</a></code> </p> <p>Other coxph tidiers: <code><a href="glance.coxph.html">glance.coxph</a>()</code>, <code><a href="tidy.coxph.html">tidy.coxph</a>()</code> </p> <p>Other survival tidiers: <code><a href="augment.survreg.html">augment.survreg</a>()</code>, <code><a href="glance.aareg.html">glance.aareg</a>()</code>, <code><a href="glance.cch.html">glance.cch</a>()</code>, <code><a href="glance.coxph.html">glance.coxph</a>()</code>, <code><a href="glance.pyears.html">glance.pyears</a>()</code>, <code><a href="glance.survdiff.html">glance.survdiff</a>()</code>, <code><a href="glance.survexp.html">glance.survexp</a>()</code>, <code><a href="glance.survfit.html">glance.survfit</a>()</code>, <code><a href="glance.survreg.html">glance.survreg</a>()</code>, <code><a href="tidy.aareg.html">tidy.aareg</a>()</code>, <code><a href="tidy.cch.html">tidy.cch</a>()</code>, <code><a href="tidy.coxph.html">tidy.coxph</a>()</code>, <code><a href="tidy.pyears.html">tidy.pyears</a>()</code>, <code><a href="tidy.survdiff.html">tidy.survdiff</a>()</code>, <code><a href="tidy.survexp.html">tidy.survexp</a>()</code>, <code><a href="tidy.survfit.html">tidy.survfit</a>()</code>, <code><a href="tidy.survreg.html">tidy.survreg</a>()</code> </p> <h3>Examples</h3> <pre> library(survival) cfit <- coxph(Surv(time, status) ~ age + sex, lung) tidy(cfit) tidy(cfit, exponentiate = TRUE) lp <- augment(cfit, lung) risks <- augment(cfit, lung, type.predict = "risk") expected <- augment(cfit, lung, type.predict = "expected") glance(cfit) # also works on clogit models resp <- levels(logan$occupation) n <- nrow(logan) indx <- rep(1:n, length(resp)) logan2 <- data.frame( logan[indx, ], id = indx, tocc = factor(rep(resp, each = n)) ) logan2$case <- (logan2$occupation == logan2$tocc) cl <- clogit(case ~ tocc + tocc:education + strata(id), logan2) tidy(cl) glance(cl) library(ggplot2) ggplot(lp, aes(age, .fitted, color = sex)) + geom_point() ggplot(risks, aes(age, .fitted, color = sex)) + geom_point() ggplot(expected, aes(time, .fitted, color = sex)) + geom_point() </pre> <hr /><div style="text-align: center;">[Package <em>broom</em> version 0.7.0 <a href="00Index.html">Index</a>]</div> </body></html>