gps.predict: Generalize Propensity Score Prediction

In fadedsoul/cdml: CDML

Description

Generalize Propensity Score Prediction

Usage

gps.predict(object, gps.method = "series", data.chunk = NULL, data.eval,
  t.grid = NULL, grid.length = 1000, verbose = FALSE,
  treatment.min = -4, treatment.max = 4, detoured = FALSE)

Arguments

object	an object from gps.method.estimation function
gps.method	selected generalized propensity score method
data.chunk	whole data, covariates are of interest. defualt = NULL
data.eval	data to be evaluated, the treatment is of interest
t.grid	alternate parameter to set up if data.chunk is NULL
grid.length	how dense the evaluation point should be like
verbose	print in console the procedure step by step
detoured	if FALSE method will be done normally; otherwise a special simulation will be examined, where CDML is chosen method, CTE is chosen model, gps.method "series" is not allowed.

Value

a vector and a matrix For the matrix elements, they stand for the predicted values of: row means given certain treatment from data.eval column means given certain covariates from data.chunk. For the vector: we list in order the predicted value as i=1,...,n for (T_i, X_i)

fadedsoul/cdml documentation built on May 14, 2019, 8:01 a.m.