Targeted maximum likelihood estimation (TMLE) of population-level causal effects under stochastic treatment regimes and related nonparametric variable importance analyses. Utilities are provided for computing the TML estimate of a counterfactual mean under a stochastically modified (continuous-valued) treatment node, under intervention policies that shift the natural value of the treatment. The methodology implemented was first described by I. Díaz and M.J. van der Laan (2013; doi: <10.1111/j.1541-0420.2011.01685.x>) and expanded upon in I. Díaz and M.J. van der Laan (2018).




GitHub repository: