In podTockom/tstmle: Data-adaptive Targeted Maximum Likelihood Estimation of Causal Effects of Interventions on Single Time Series

knitr::opts_chunk$set(
  collapse = TRUE,
  comment = "#>",
  fig.path = "README-"
)

R/`tstmle3`

Data-adaptive Estimation and Inference for Causal Effects with a Single Time Series

Authors: Ivana Malenica

What's `tstmle3`?

The tstmle3 implements robust estimation and provides inference for data-dependent causal effects based observing a single time series. It's an adapter/extension R package in the tlverse ecosystem.

Consider the case where one observes a single time-series, denoted as a single sequence of dependent random variables $O(1), \dots O(N)$ where each $O(t)$ with $t \in {1, \dots ,N}$ takes values in $\mathbf{R}^p$. Further, we assume that at each time $t$, we have a chronological order of the treatment or exposure $A(t)$, outcome of interest $Y(t)$, and possibly other covariates $W(t)$. While studying time-series data, one might be interested in what the conditional mean of the outcome would have been had we intervened on one or more of the treatment nodes in the observed time-series.

The tstmle3 package focuses on a class of statistical target parameters defined as the average over time $t$ of context-specific pathwise differentiable target parameters of the conditional distribution of the time-series [@c2]. In particular, it implements several context-specific causal parameters that can be estimated in a double robust manner and therefore fully utilize the sequential randomization.

In particular, tstmle3 implements few different context-specific parameters:

Average over time of context-specific ATE of a single time point intervention.
Average over time of context-specific TSM of a single time point intervention.

Here, initial estimation is based on the sl3 package, which constructs ensemble models with proven optimality properties for time-series data [@c3].

Installation

You can install a stable release of tstmle from GitHub via devtools with:

devtools::install_github("imalenica/tstmle3")

Note that in order to run tstmle you will also need sl3 and tmle3:

devtools::install_github("tlverse/sl3")
devtools::install_github("tlverse/tmle3")

Issues

If you encounter any bugs or have any specific feature requests, please file an issue.

Citation

After using the tstmle3 R package, please cite the following:

@software{malenica2022tstmle3,
      author = {Malenica, Ivana and {van der Laan}, Mark J},
      title = {{tstmle3}: Context-Specific Targeted Learning for time-series},
      year  = {2022},
      doi = {},
      url = {https://github.com/imalenica/tstmle3},
      note = {R package version 1.0.0}
    }

License

The contents of this repository are distributed under the MIT license. See below for details:

The MIT License (MIT)

Copyright (c) 2022-2023

Permission is hereby granted, free of charge, to any person obtaining a copy
of this software and associated documentation files (the "Software"), to deal
in the Software without restriction, including without limitation the rights
to use, copy, modify, merge, publish, distribute, sublicense, and/or sell
copies of the Software, and to permit persons to whom the Software is
furnished to do so, subject to the following conditions:

The above copyright notice and this permission notice shall be included in all
copies or substantial portions of the Software.

THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
SOFTWARE.