Gradient boosted propensity score estimation
ps calculates propensity scores using gradient boosted logistic regression and diagnoses the resulting propensity scores using a variety of methods
ps( formula = formula(data), data, n.trees = 10000, interaction.depth = 3, shrinkage = 0.01, bag.fraction = 1, n.minobsinnode = 10, perm.test.iters = 0, print.level = 2, verbose = TRUE, estimand = "ATE", stop.method = c("ks.mean", "es.mean"), sampw = NULL, version = "gbm", ks.exact = NULL, n.keep = 1, n.grid = 25, keep.data = TRUE, ... )
formula: An object of class formula: a symbolic description of the propensity score model to be fit with the treatment indicator on the left side of the formula and the potential confounding variables on the right side.
data: A dataset that includes the treatment indicator as well as the potential confounding variables.
n.trees: Number of gbm iterations passed on to gbm::gbm(). Default: 10000.
interaction.depth: A positive integer denoting the tree depth used in gradient boosting. Default: 3.
shrinkage: A numeric value between 0 and 1 denoting the learning rate. See gbm for more details. Default: 0.01.
bag.fraction: A numeric value between 0 and 1 denoting the fraction of the observations randomly selected in each iteration of the gradient boosting algorithm to propose the next tree. See gbm for more details. Default: 1.0.
n.minobsinnode: An integer specifying the minimum number of observations in the terminal nodes of the trees used in the gradient boosting. See gbm for more details. Default: 10.
perm.test.iters: A non-negative integer giving the number of iterations of the permutation test for the KS statistic. If perm.test.iters=0
then the function returns an analytic approximation to the p-value. Setting perm.test.iters=200 will yield precision to within 3% if the true p-value is 0.05. Use perm.test.iters=500 to be within 2%. Default: 0.
print.level: The amount of detail to print to the screen. Default: 2.
verbose: If TRUE, lots of information will be printed to monitor the the progress of the fitting. Default: TRUE.
estimand: "ATE" (average treatment effect) or "ATT" (average treatment effect on the treated) : the causal effect of interest. ATE estimates the change in the outcome if the treatment were applied to the entire population versus if the control were applied to the entire population. ATT estimates the analogous effect, averaging only over the treated population. Default: "ATE".
stop.method: A method or methods of measuring and summarizing balance across pretreatment variables. Current options are ks.mean, ks.max, es.mean, and es.max. ks refers to the Kolmogorov-Smirnov statistic and es refers to standardized effect size. These are summarized across the pretreatment variables by either the maximum (.max) or the mean (.mean). Default: c("ks.mean", "es.mean").
sampw: Optional sampling weights.
version: "gbm", "xgboost", or "legacy", indicating which version of the twang package to use.
"gbm": uses gradient boosting from the gbm package,"xgboost": uses gradient boosting from the xgboost package, and"legacy": uses the prior implementation of the ps function.Default: "gbm".
ks.exact: NULL or a logical indicating whether the Kolmogorov-Smirnov p-value should be based on an approximation of exact distribution from an unweighted two-sample Kolmogorov-Smirnov test. If NULL, the approximation based on the exact distribution is computed if the product of the effective sample sizes is less than 10,000. Otherwise, an approximation based on the asymptotic distribution is used. Warning: setting ks.exact = TRUE will add substantial computation time for larger sample sizes. Default: NULL.
n.keep: A numeric variable indicating the algorithm should only consider every n.keep-th iteration of the propensity score model and optimize balance over this set instead of all iterations. Default: 1.
n.grid: A numeric variable that sets the grid size for an initial search of the region most likely to minimize the stop.method. A value of n.grid=50 uses a 50 point grid from 1:n.trees. It finds the minimum, say at grid point 35. It then looks for the actual minimum between grid points 34 and 36. If specified with n.keep>1, n.grid
corresponds to a grid of points on the kept iterations as defined by n.keep. Default: 25.
keep.data: A logical variable indicating whether or not the data is saved in the resulting ps object. Default: TRUE.
...: Additional arguments that are passed to ps function.
Returns an object of class ps, a list containing
gbm.obj: The returned gbm or xgboost object.
treat: The vector of treatment indicators.
treat.var: The treatment variable.
desc: A list containing balance tables for each method selected in stop.methods. Includes a component for the unweighted analysis names unw . Each desc component includes a list with the following components
- **`ess`**: The effective sample size of the control group.
- **`n.treat`**: The number of subjects in the treatment group.
- **`n.ctrl`**: The number of subjects in the control group.
- **`max.es`**: The largest effect size across the covariates.
- **`mean.es`**: The mean absolute effect size.
- **`max.ks`**: The largest KS statistic across the covariates.
- **`mean.ks`**: The average KS statistic across the covariates.
- **`bal.tab`**: a (potentially large) table summarizing the quality of the weights for equalizing the distribution of features across the two groups. This table is best extracted using the `bal.table` method. See the help for `bal.table` for details on the table's contents.
- **`n.trees`**: The estimated optimal number of gradient boosted iterations to optimize the loss function for the associated `stop.methods`.
- **`ps`**: a data frame containing the estimated propensity scores. Each column is associated with one of the methods selected in `stop.methods`.
- **`w`**: a data frame containing the propensity score weights. Each column is associated with one of the methods selected in `stop.methods`.
- **`estimand`**: The estimand of interest (ATT or ATE).
datestamp: Records the date of the analysis.
parameters: Saves the ps call.
alerts: Text containing any warnings accumulated during the estimation.
iters: A sequence of iterations used in the GBM fits used by plot function.
balance: The balance measures for the pretreatment covariates used in plotting, with a column for each stop.method.
balance.ks: The KS balance measures for the pretreatment covariates used in plotting, with a column for each covariate.
balance.es: The standard differences for the pretreatment covariates used in plotting, with a column for each covariate.
ks: The KS balance measures for the pretreatment covariates on a finer grid, with a column for each covariate.
es: The standard differences for the pretreatment covariates on a finer grid, with a column for each covariate.
n.trees: Maximum number of trees considered in GBM fit.
data: Data as specified in the data argument.
For user more comfortable with the options of xgboost::xgboost(), the options for ps controlling the behavior of the gradient boosting algorithm can be specified using the xgboost naming scheme. This includes nrounds, max_depth, eta, and subsample. In addition, the list of parameters passed to xgboost can be specified with params.
Note that unlike earlier versions of twang, the plotting functions are no longer included in the ps function. See plot for details of the plots.
Dan McCaffrey, G. Ridgeway, Andrew Morral (2004). "Propensity Score Estimation with Boosted Regression for Evaluating Adolescent Substance Abuse Treatment", Psychological Methods 9(4):403-425.
gbm, xgboost, plot, bal.table
Useful links