How deeply does Causal Inference for Data Science cover s-learner?

Defines a four-step prediction procedure and notes unrealistic results as a data-dependent limitation.

S-learner

As explained in Causal Inference for Data Science

An S-learner is a causal estimation algorithm that fits one outcome model containing the treatment and confounders and predicts outcomes under alternative treatment values.

An S-learner is a simple causal estimation method: train one machine-learning model using treatment X and confounders Z, then predict outcomes under alternative treatment values. It applies these predictions within the adjustment formula to estimate quantities such as E[Y|do(X = x)] and the average treatment effect (ATE). section 4.2.4

Estimate an intervention by following these steps:

Train a model f on historical observations containing treatment X, outcome Y, and all confounders Z1, …, Zp. section 4.2.4
Use a classifier for categorical outcomes or a regressor for continuous outcomes. section 4.2.4
For every observation, replace its observed treatment with the value of interest, such as X = 1 and X = 0, while retaining its confounders. section 4.2.4
Average the predicted outcomes to estimate the ATE; the single-model approach is simple but can produce unrealistic results depending on the data. section 4.2.4

How Causal Inference for Data Science covers it

Cover of Causal Inference for Data Science

liveBook manning.com

Read Causal Inference for Data Science on liveBook Get the book at manning.com

Related concepts

causal adjustmentuses · Causal adjustment estimates a causal effect by conditioning on variables that block noncausal pathways and averaging over their target-population distribution.
T-learneraddressed by · A T-learner is a causal inference method that fits separate outcome models for each treatment level and contrasts their predictions.
double machine learningcontrasts with · Double machine learning is a causal-effect estimation method that uses machine-learning nuisance models with orthogonalized estimation to reduce bias from nuisance-function errors.
decision treeuses · A decision tree is a branching structure that represents sequential tests or choices, with paths leading to outcomes, predictions, or actions.

This explainer was generated by Manning's knowledge graph from the text of Causal Inference for Data Science. The section links point to the parts of the book it draws on.