SigmaMonstR / ate

ATE

A prototype R library to calculate ML-based treatment effects per Varian (2014).

Example of how to use autoTrain

First, load in the code.

  source("https://raw.githubusercontent.com/SigmaMonstR/ate/master/script/autoTrain.R")

Then simulate some fake data

  #Set up data
    n <- 10000
    df <- data.frame(id = 1:n,
                     time = round((1:n)/100),
                     y = as.numeric(1:n + rnorm(n,30,300)),
                     x1 = rnorm(n, 100,100),
                     x2 = rnorm(n, 200,100),
                     x3 = runif(n)*(1:n),
                     x4 = rnorm(n, 100,100),
                     x5 = rnorm(n, 200,100),
                     x6 = runif(n)*(1:n))
    df <- df[complete.cases(df),]
    tt <- runif(nrow(df)) >= 0.6

Set up a parameter file:

    args <- list(formula = as.formula("y ~ x1 + x2 + x3 + x4 + x5 + x6"),
                 data = df,
                 treatment = tt,
                 window.type = 2,
                 w.size = 10,
                 w.factor = 5,
                 algo = 2,
                 run.var = "time",
                 kval = 10,
                 rdd.thresh = 50)

Then apply:

    out <- autoTrain(args)