Comments (1)
@athammad, thanks for your question. Unfortunately, the effect you appear to describe does not align with the target parameter(s) whose estimators we implement in this package. Note that the effects we describe in https://arxiv.org/abs/1901.02776 are stochastic direct and indirect effects that are based on a joint stochastic intervention that generates counterfactuals based on modifying the exposure but drawing the mediator(s) based on the (true) natural distribution of the exposure.
In order to accommodate multiple mediators, we include a re-parameterization that substitutes a mediation mechanism (e.g., f(Z | A, W)) for an exposure mechanism conditional on the mediators (e.g., f(A | Z, W)), thereby allowing for multiple mediators to be easily considered while avoiding the complications that would arise from estimating a joint conditional distribution over multiple mediators. Thus, the effects you have estimated in your example correspond to the direct and indirect effects for multiplying the odds of exposure by 3 while drawing the mediators (Z and ZZ) from the natural distribution of the exposure (prior to multiplying it's odds by 3). The effect decomposition we provide does not partition the effects through individual mediators, as that would be a different causal inference problem entirely.
from medshift.
Related Issues (20)
- test indexing approaches HOT 1
- TODO HOT 1
- clearer documentation and naming
- missing outcome support
- Arbitrary fold structures for one-step estimator HOT 1
- Scaling transformation of outcome variable
- TML estimator for binary interventions HOT 2
- Support for continuous-valued exposures
- Check that CV-folds is greater than 1
- testing Monte Carlo integration
- Nuisance parameter phi should use training data for one-step HOT 1
- weight stabilization HOT 1
- Support for observation-level IDs HOT 1
- stochastic interventional (in)direct effects
- Functionality for true continuous treatment HOT 1
- substitution estimator should use counterfactuals HOT 1
- problems in efficient estimator HOT 1
- utility function for IP weights HOT 1
- utility function for Dzw/substitution HOT 1
Recommend Projects
-
React
A declarative, efficient, and flexible JavaScript library for building user interfaces.
-
Vue.js
🖖 Vue.js is a progressive, incrementally-adoptable JavaScript framework for building UI on the web.
-
Typescript
TypeScript is a superset of JavaScript that compiles to clean JavaScript output.
-
TensorFlow
An Open Source Machine Learning Framework for Everyone
-
Django
The Web framework for perfectionists with deadlines.
-
Laravel
A PHP framework for web artisans
-
D3
Bring data to life with SVG, Canvas and HTML. 📊📈🎉
-
Recommend Topics
-
javascript
JavaScript (JS) is a lightweight interpreted programming language with first-class functions.
-
web
Some thing interesting about web. New door for the world.
-
server
A server is a program made to process requests and deliver data to clients.
-
Machine learning
Machine learning is a way of modeling and interpreting data that allows a piece of software to respond intelligently.
-
Visualization
Some thing interesting about visualization, use data art
-
Game
Some thing interesting about game, make everyone happy.
Recommend Org
-
Facebook
We are working to build community through open source technology. NB: members must have two-factor auth.
-
Microsoft
Open source projects and samples from Microsoft.
-
Google
Google ❤️ Open Source for everyone.
-
Alibaba
Alibaba Open Source for everyone
-
D3
Data-Driven Documents codes.
-
Tencent
China tencent open source team.
from medshift.