calc_mixtures_ate: Calculate the ATE for each mixture rule

In blind-contours/CVtreeMLE: Cross Validated Decision Trees with Targeted Maximum Likelihood Estimation

Calculate the ATE for each mixture rule

Description

Aggregate mixture rules found across the folds that have the same variables. For each rule extract the relevant nuisance parameter data calculated in the folds. Given the validation data estimates across the folds, for each tree do a TMLE update step to target the average treatment effect. Update the initial counterfactuals, calculate the influence curve and using the influence curve calculate variance estimates and p-values.

Usage

calc_mixtures_ate(input_mix_rules, input_mix_data, y, n_folds)

Arguments

input_mix_rules	List of dataframes of rules found for a mixture across the folds
input_mix_data	Nuisance parameter data for mixture rules found
y	Outcome variable across the folds
n_folds	Number of folds used in cross-validation across all the folds

Value

A list with mixture analysis results which includes:

• results: A data frame with variable threshold combinations on the rows and ATE, variance and consistency estimates on the columns.

• group_list: A list of rule combinations found in the ensemble decision tree model grouped by variable sets in the rules and directions of the coefficient in the linear model. Also provided is the fold the rule was found and the RMSE of the model which used the rule.

• mixture_data_list: A list of data frames which houses the data for each mixture rule evaluated as an exposure, the baseline covariates, outcome, and nuisance parameter estimates for the respective rule.

blind-contours/CVtreeMLE documentation built on June 22, 2024, 8:53 p.m.