Can we train machine learning methods to outperform the high dimensional propensity score algorithm?
Our results suggest that, when the data analysis is done with epidemiologic principles in mind, machine learning methods perform as well as the high-dimensional propensity score algorithm.
Does the estimation of the propensity score by machine learning improve matching estimation?
Although off-the-shelf machine learning methods have many well-documented advantages in prediction and classification, it is not obvious that using them for propensity score estimation in a matching framework will improve the estimation of causal effects.
What is propensity score in machine learning?
The propensity score is the probability of receiving a treatment conditional on a set of observed covariates [1]. At each value of the propensity score, the distribution of observed covariates is the same across treatment groups. Propensity scores are generally estimated using logistic regression.
How can machine learning be used to calculate propensity score?
There have been many alternatives developed that use machine learning methods to flexibly model the propensity score. The two most effective from what I’ve seen have been Bayesian Additive Regression Trees (BART; Hill, 2011; applied to propensity score modeling Hill et al., 2011) and SuperLearner (Pirrachio et al., 2015).
How is machine learning used in causal inference?
Most causal inference algorithms usually have some machine learning core — a statistical model that predicts the outcome or treatment assignment. Once a mapping between features to targets is obtained, causal models can then have various ways to indirectly apply those statistical predictions to obtain a causal estimate.
Which is the best method for machine learning?
SuperLearner is a stacking method that allows you to supply many different machine learning methods and it either picks the best one or takes an optimally weighted combination of them. If any of the machine learning methods approximate the true model, SuperLearner will perform as well or better than the best (asymptotically).
Why is the propensity score so important in inference?
The first approach relies on the finding that balancing on a well-formed propensity score balances all pre-treatment covariates fully (i.e., their entire joint distribution). This is what Rosenbaum & Rubin (1983) discovered and why the propensity score has become so important.